Jukuri, open repository of the Natural Resources Institute Finland (Luke) All material supplied via Jukuri is protected by copyright and other intellectual property rights. Duplication or sale, in electronic or print form, of any part of the repository collections is prohibited. Making electronic or print copies of the material is permitted only for your own personal use or for educational purposes. For other purposes, this article may be used in accordance with the publisher’s terms. There may be differences between this version and the publisher’s version. You are advised to cite the publisher’s version. This is an electronic reprint of the original article. This reprint may differ from the original in pagination and typographic detail. Author(s): Andreas Otterbeck, Andreas Lindén Title: Temporal increase in migratoriness and increasing male bias among residents in partially migrating Swedish sparrowhawks Accipiter nisus Year: 2024 Version: Published version Copyright: The Author(s) 2024 Rights: CC BY 4.0 Rights url: https://creativecommons.org/licenses/by/4.0/ Please cite the original version: Otterbeck, A., & Lindén, A. (2024). Temporal increase in migratoriness and increasing male bias among residents in partially migrating Swedish sparrowhawks Accipiter nisus. Ornis Fennica, 101(4), 116–130. https://doi.org/10.51812/of.122172 (Original work published December 13, 2024). Ornis Fennica 101: 116–130. 2024 1. Introduction Climate responses in bird migration have received much attention, with a particular focus on changes in migration phenology (Rubolini et al. 2007, Usui et al. 2017, Lehikoinen et al. 2019) and how insufficient responses lead to temporal mismatches with seasonal resource peaks (Visser Partial migrants have populations consisting of both migratory and resident individuals. These migrants and residents experience unequal ecological conditions during winter and the underlying factors driving their decision to stay on their breeding grounds or to migrate remain debated—both from the viewpoint of populations and individuals. Here, we studied partial migration in a small raptor, the Eurasian Sparrowhawk (Accipiter nisus), from two different but interconnected perspectives: 1) explaining the patterns and variation in the ratio of migrants to residents (migratoriness) at the population level and 2) revealing how age and sex may affect the individual decision to be migratory or resident. We used citizen observation data over four decades to explore the temporal and spatial variation in the age and sex ratio of wintering resident sparrowhawks in Sweden. We found that the migratoriness unexpectedly increased with higher annual temperatures and showed long-term trend across the study period. Also, this migrant-to- resident ratio increased with smaller winter prey abundance. The average winter sex ratio was male-biased and became increasingly so over the years. We suggest that residency benefits territory-establishing males as early presence gives a competitive advantage in obtaining high-quality territories. Moreover, the distribution of overwintering individuals (regardless of sex) moved gradually northwards as the winter progressed, suggesting that smaller-scale migration occurs among the resident fraction of the population. These results provide suggestions for the underlying drivers and regulation of partial migration. A. Otterbeck, The Helsinki Lab of Ornithology, Finnish Museum of Natural History, University of Helsinki, Finland & Novia University of Applied Sciences, Raseborgsvägen 9, FI-10600, Ekenäs, Finland A. Lindén, Natural Resources Institute Finland (Luke), P.O. Box 2, FI-00791, Helsinki, Finland * Corresponding author’s e-mail: andreas.otterbeck@helsinki.fi Received 10 October 2022, accepted 11 November 2024 Temporal increase in migratoriness and increasing male bias among residents in partially migrating Swedish sparrowhawks Accipiter nisus Andreas Otterbeck* & Andreas Lindén 117 ORNIS FENNICA Vol.101, 2024 & Gienapp 2019). Migration enables individ- uals to utilize seasonal resources, but there is a growing interest in why some birds migrate while others do not. In partial migrants, a single population contains both migratory and resident individuals, which is fairly common among short-distance migrants (Lack 1943, 1944, Terrill & Able 1988, Lundberg 1988, Chapman et al. 2011). Rising temperatures may tip the balance of whether being migratory is a superior strategy for individual life history, which may alter the composition of migrants and residents within and across partially migratory populations. Such responses have classically received little attention compared to other spatial and temporal aspects of migration, e.g., the timing of arrival in spring (Lundberg 1988, Newton 2008, Chapman et al. 2011) despite its potential to cause rapid, large, and unpredictable effects on spatial distri- bution, population dynamics, and life history. The average migratoriness has been shown to increase in populations toward areas with harsher winter environments, such as higher latitudes in the northern hemisphere (Main 2002, Newton 2008, Boyle et al. 2010, Somveille et al. 2013, Ambrosini et al. 2016) and altitudes (Boyle et al. 2010, Lundblad & Conway 2020). It has been hy- pothesized that partial migrants may step towards year-round residency, as a way to adapt to climate warming, when the survival prospects of overwin- tering near breeding latitudes improve (Berthold 1996, 2001, Pulido & Berthold 2010, Chapman et al. 2011, Meller et al. 2016). While the spatial patterns in migratoriness may result from local adaption over time, and for a few species show temporal responses to annual temperature, there are limited examples showing how migratoriness could directionally divert over time, currently suggesting heterogeneous trends across species (Nilsson et al. 2006, Van Vliet et al. 2009, Meller et al. 2016). While the individual migration decision is binary, the migratory propensity is continuous and heritable (Berthold & Querner 1982, Berthold 1988, 1999, 2001, Biebach 1983, Pulido & Berthold 2010), but to a variable degree plastically modified by environmental compo- nents (Able & Belthoff 1998). Besides long-term selection by temperature, the fitness prospects by residency may be additionally affected by factors such as food abundance (Lindén et al. 2011, Meller et al. 2016), extreme weather (Acker et al. 2021) and density dependence (Kokko & Lundberg 2001, Lundberg 1988, Meller et al. 2016), potentially at a shorter temporal scale. Thus, the drivers for migratory polymorphism are likely conditional; key individual costs and benefits from either migratory or resident pheno- types remain controversial, and generalizations are yet difficult (Chapman et al. 2011). Heterogeneous responses in migratoriness suggest that the fitness of migrants and residents is individual and conditional in space and time. Within a population, the individual pressure to migrate should be unequal between individuals to maintain both strategies (Pulido & Berthold 2010, but see Pulido 2011, de Zoeten & Pulido 2020). In the temperate region, the migratory and resident fraction may experience very different ecological conditions and life history, and the underlying factors driving this polymorphic trait remain debated both for populations and at an individual level. Recent literature has suggested both scenarios with a fitness benefit for the resident fraction (i.e., Grist et al. 2017, Buchan et al. 2020) and the migratory fraction of a pop- ulation (Zúñiga et al. 2017, Acker et al. 2021), underscoring a lack of uniformity in fitness advantage between the two migratory pheno- types. Also, there is a discrepancy in whether such a benefit is primarily driven by survival (Buchan et al. 2020, Zúñiga et al. 2017) which would likely affect differently among individuals and areas. Thus, the actual costs and gains an individual faces as a migrant or a resident are likely conditional, depending on pre-existing traits and capabilities, such as migration distance, capability of survival in the cold, and subsequent breeding success. What phenotypic traits facilitate migration or residency as the superior strategy remains contro- versial as well, with competing hypotheses having suffered from overlapping predictions (Chapman et al. 2011). The most central theories predicting individual migratory propensity concern sex and age, and we here explore three such hypotheses, with predictions for which individuals show higher/lower migratory propensity. First, the arrival time hypothesis states that early spring presence at the breeding grounds gives a competitive advantage for high-quality Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 118 territories, favouring residency for the territory establishing sex (Ketterson & Nolan 1976), usually males (e.g. Schwabl 1983). This, however, assumes that: a) high-quality territories are subject to intra-specific competition, b) resident individuals are best able to acquire such territories by arriving early and c) such acqui- sition gives a fitness advantage through, e.g., improved fecundity and survival. Secondly, the body size hypothesis is based on the thermoregulatory superiority of larger body size. Reduced surface area to volume ratio decreases heat loss and body size is positively correlated with basal metabolism (Daan et al. 1990). This predicts larger residents (Ketterson & Nolan 1976, Belthoff & Gauthreaux 1991, Chapman et al. 2011, Lehikoinen 2011) which for most bird species with sexual size dimorphism would be males (Dunning 2008). Third, the competitive release hypothesis (also referred to as dominance hypothesis) states that if the available winter habitat for residents is restricted, density-dependent competition may occur. If so, dominant individuals (usually the larger individuals) should gain a competitive advantage forcing subordinates (usually females and juveniles) to migrate (Ketterson & Nolan 1976, Gauthreaux 1982, Lundberg 1985, Smith & Nilsson 1987). These three hypotheses classi- cally overlap in their predictions, suggesting that males are more often resident (Chapman et al. 2011) and part of the literature on this topic stems from obligate migrants, using migration distance as the response variable instead of the propensity to migrate at all. Our aim here is divided into two interconnect- ed parts, focusing on the Eurasian Sparrowhawk (Accipiter nisus) as a study system. Firstly, to approximate changes in the migratoriness on a population level, we study the temporal trends and annual variation in the migrant to resident ratio (hereafter ‘MRR’), relating it to annual tem- perature anomalies during the migration period, and food availability, measured as the annual winter prey abundance. We expect that lower temperatures and lower food availability drive more individuals to migrate. We also predict that sparrowhawks have become less migratory as a long-term response to climate warming. Secondly, we study how sex and age may conditionally affect the costs and benefits of being a resident or migrant, and evaluate the evidence for the three aforementioned hypoth- eses. Studying the sparrowhawk enables us to untangle overlapping predictions due to the species’ reversed sexual size dimorphism. A male- biased sex ratio in winter would support the arrival time hypothesis while more females would suggest the body size hypothesis or the competitive release hypothesis—the latter would be suggested by resident adults. Besides testing the predictions of these hypotheses, we also explore the sex- and age-specific spatio-temporal distribution patterns, such as the latitudinal distri- bution and long-term temporal change to obtain a more complete picture of their consistency. In all sets, we explore through model selection a possible connection between our study parts 1 and 2; whether the annual variation in the MRR co-varies with the ratio of sex, age or latitude distribution. Such a pattern would further clarify how certain individuals are more likely to migrate when the pressure to migrate increases. 2. Materials and methods 2.1. Study species We studied the Eurasian Sparrowhawk (Accipiter nisus); a small-sized, partially migratory raptor with an extreme reversed size dimorphism, which is also the most common raptor in Fennoscandia (Newton 1986). Females can be more than twice as heavy as males (Dunning 2008), likely providing better thermal resilience in cold winters while leaving males physically subordinate (Newton 1986). The parental effort is skewed; the male establishes the territory and feeds the female, the chicks and himself throughout a significant part of the breeding season (Newton 1986). The quality of territo- ries varies (Newton 1986), and the local prey availability and exploitation can affect breeding productivity (Otterbeck et al. 2015). Thus, acquiring high-quality territories is under strong competition among territory-establishing males. Sex separation and age determination (juvenile/ adult) are fairly easy in the field for more experi- enced observers. 119 ORNIS FENNICA Vol.101, 2024 2.2. Study area and weather data Our study area, Sweden, spans over a large latitu- dinal range and has a long tradition of surveying and reporting both summer and winter bird abundances. Moreover, there is a well-monitored migration hotspot in the south at Falsterbo where a substantial part of migratory sparrowhawks will pass on their journey south. Assuming facultative migration in at least part of the population, partial migrants may adapt their migratory phenotype to prevailing environmental conditions before or during their migration period. To measure such inter-annual variation, we calculated an annual average late autumn temperature (from October 1 to November 30) using temperature data from 7 districts evenly scattered from south to north from 1975 to 2013. The data were obtained from the Swedish Meteorological Institute. This late-autumn period can only affect late migrants, mainly adults, who haven’t yet migrated when encountering a relevant drop in temperature. Thus, this period evades the peak of juvenile migration in August–September, during which temperature is not likely to affect the migration decision. 2.3. Avian abundance data 2.3.1. Migration survey and wintering abundance indices We used annual total migration counts of spar- rowhawks from Falsterbo Bird Observatory. The observatory is situated on a peninsula at the southwestern tip of Sweden (55°23 ´ N, 12°5 ´ E) which acts as a major migration funnel for autumn migration out of Fennoscandia. Although migration occurs along several other pathways, raptors are thought to be especially well represented at this site (Karlsson 2004). Even if not all sparrowhawks are observed, we here suppose that the annual totals at Falsterbo are approximately proportional to the numbers migrating out of Sweden. Here the peak migration of juvenile sparrowhawks occurs in September, and the adults in October. Juveniles comprise the largest demographic group of recorded migrants (Kjellén 1992). The official survey switched observer in 2001, along with some adjusted methodology. To account for possible effects on the time series from this, we modelled the effect of the observer shift in 2001 using a generalized additive model (GAM) to test and correct for any disruption in the counts before and after the switch (Supplementary Material S1, Fig. S1). The annual winter censuses were part of the organized annual Swedish bird census monitoring scheme (Green & Lindström 2014). The survey was conducted by amateur ornithologists from 1975 to 2016 based on 248–644 point census routes per winter. Every route is observer-chosen and has 20 evenly spread points, each of which is surveyed for 5 minutes. The route is repeated annually whenever possible. The minimum sepa- ration between the points was 300 meters in dense forests and 300–400 meters in open landscapes. We used published annual winter abundance indexes from Green & Lindström (2014), which are based on the statistical software TRIM (Pannekoek & van Strien 2005). 2.3.2. Winter abundance of resident prey species To assess the annual variation in the winter prey abundance for sparrowhawks, we calculated the annual weighted averages of the winter popu- lation abundances of the 10 most central prey species for sparrowhawks. The scaled population indices of each species were weighted with the average number of birds observed annually, before averaging. Finally, the annual prey indices (‘Win.prey’) were scaled by dividing the result by the average over all years, so that the mean of the resulting index is one (Supplementary Material Table S2). 2.3.3. Population structure and latitudinal distribution of wintering individuals We analysed winter observations of sparrow- hawks in Sweden during 1975–2016, extracted from the portal (www.artportalen.se). We extracted observations spanning from day-of- year 305 in late autumn to day 91 the following spring (in non-leap years: November 1 – April 1). This period should fairly well avoid the arrival of Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 120 spring migrants. Each winter season was named after the year starting in January. We described the day of winter (‘Dow’) where January 1 is day 1 and the previous year’s days get smaller values than that (e.g. December 31 is day 0, December 30 is day –1, and so on). We only included observations where the sex had been determined and reported—reducing the risk of including lower-quality observations. Citizen science data may contain high levels of noise, reflecting both the observer’s skills and reporting behaviour, but this subset likely provided at least a notable reduction in the most obvious sources of bias. There is a difference in the difficulty of aging males and females in the field, and the age of females was less frequently determined compared to males in the data (X2-test: X2= 156.13, df = 2, P < 0.001). This is expected, as the adult female plumage is closer to juveniles compared to that of males. The total sample size with known sex was 17,180, dropping to 5,578 when adding the criteria of known age (2 levels: 1 cy and adult) (Table 1). Another potential confusion is between female sparrowhawks and male goshawks (Accipiter gentilis), although the goshawk is approximately twice the weight. Occasional misidentifications would likely add to the number of sparrowhawk females in our data. We assume that long-term changes in the average identification skill/ability of the observers are merely reflected in the sample size, and that the criteria or threshold for identifying the age or sex (or species) have remained constant, and will therefore mainly be reflected in the estimated average (intercept). 2.4. Statistical analyses All the statistical analyses were conducted in R version 3.5.1 (R Core Team 2018). For fitting GAMs we used packages ‘mgcv’ (Wood 2017) and ‘gamm4’ (Wood & Scheipl 2020). Table 1. Brief explanations of the variables used in this study. The information listed includes the type of variable, the range of values (or factor levels), and usage of the variables in the analysis. Zero-centred variables have been pre- processed by subtracting the average from each observation, so that the new mean is zero. Variable name Explanation Type Range (unscaled) Population level analysis Migration Migration count at Falsterbo numeric 8.32–10.72 Win.pop Population index winter numeric 0.33–1.67 ln.MRR ln(Migration / Win.pop) numeric (response) (–2.16)–1.16 Win.prey Abundance sum of 10 prey species numeric 0.43–3.99 Temp.c Zero-centered temperature Oct & Nov numeric (–3.34)–3.32 Year.c Zero-centered year, for temporal trend numeric (–20.5)–20.50 Sex and age ratios of residents is.male Individual sex (1 = male, 0 = female) binary (response) 1 or 0 is.adult Individual age (1 = older, 0 = first year) binary (response) 1 or 0 Lat Latitude of the observation numeric (response) 55.34–68.36 Dow Day of the winter centralized on 1.jan numeric (–60)–90 Year.f Year (factor), for annual variation factor, 39 levels “1975”–“2013” ln.MRR.c Zero-centred ln migrant to resident ratio numeric (–2.07)–1.24 Year.c Zero-centred year, for temporal trend numeric (–31.99)–6.01 121 ORNIS FENNICA Vol.101, 2024 2.4.1. The ratio of migrants to residents (MRR) We studied annual variation in migratoriness, by examining the ratio of observed migrants to wintering individuals. To do so, we apply a multiple regression model with the natural logarithm of the migrant to wintering ratio (‘ln.MRR’), denoted as ln ( ‘Migration’ / ‘Win. pop’), as the response variable. As the explana- tory variables we applied the three zero-centred variables ‘Year.c’, ‘Temp.c’ and ‘Win.prey.c’ (Table 1), to accommodate the long-term trend over 40 years, the effects of late-autumn tempera- ture, and annual index of winter prey availability, respectively. Notice that this approach detects any effects on the number of migrating or wintering individuals, or both simultaneously. To further explore whether the observed results were driven by the migration data, wintering data, or the con- tribution of both, we also fitted the model to the natural logarithms of ‘Migration’ and ‘Win.pop’ only. 2.4.2. Sex ratio, age ratio and latitude of resident observations We applied model selection using the Akaike Information Criterion creating sets of general- ized additive models (GAM) for studying the 1) sex ratio 2) age ratio and 3) reported latitude (Supplementary Material Table S3). For sex and age (cases 1–2) we apply a logit link function and a binomial error distribution, while for latitude (case 3) we use an identity link and normal error distribution. For each hypothesis, we set up three groups of model complexity concerning the explanatory variables; a) zero-models with no covariates, b) the covariate ‘Year.c’ coding for a temporal trend, and c) the covariate ‘ln.MRR.c’ which in this case represents the log of the annual migrant to resident ratio. As the ‘Year.c’ and ‘ln.MRR.c’ showed a high positive correlation (r = 0.7), these could not co-occur in the same model, but were regarded as alternative explanations in competing models. However, as ‘Year.c’ effectively rep- resents a straight line, we expected the model with ‘ln.MRR.c’ to show a superior negative likelihood for ‘ln.MRR.c’ to be an informative parameter sensu Arnold (2010). In all models, we modelled within-seasonal patterns by fitting ‘Dow’ (continuous variable indicating date of the winter season) with a smoothing function, using the default options of the ‘gam’ function in the ‘mgcv’ package (thin-plate spline as smoothing basis, max. df = 9). There are two main reasons for doing this; 1) it makes the definition of start and end of the wintering season less critical and 2) helps to safeguard against the pitfalls of the possible biased citizen science data. Our logic is that if, for instance, adult males are truly more common during the winter, we would likely see a changing sex- and age-ratio with a mid-winter peak. In all models, we included year as a the factor variable of year ‘Year.f’ as a random effect that captures the unex- plained of annual variation. The binary variable ‘is.male’ was identified as ‘false’ = female and ‘true’ = male (Table 1) and we assume a 1:1 sex ratio in the population, but acknowledge that we do not know the actual sex ratio. However, an equal sex ratio has been reported for broods (e.g. Risch & Brinkhof 2002). We grouped the age of the individuals into the variable ‘is.adult’ with the two categories ‘false’ (i.e. juveniles: 1 cy or 2 cy after new year) and ‘true’ (i.e. adults: 2 cy before new year, or older than 2 cy) (Table 1). While we applied the results from package ‘mgcv’ for model selection, we re-run the best models of each hypothesis using the ‘gamm4’ package to obtain the estimated standard deviations of the random effect. Also other effects presented and illustrated in figures are based on the refitted best models. 3. Results 3.1. The migrant-to-resident ratio The migrant-to-resident ratio showed a positive annual relationship with autumn temperature (estimate: 0.065, SE = 0.031, t = 2.10, P = 0.042) and a negative relationship with the annual abundance of the most relevant prey species (estimate: –0.205, SE = 0.083, t = –2.45, P = 0.019). Further, there was a positive partial trend Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 122 in the migrant-to-resident ratio across the study period (estimate: 0.054, SE = 0.004, t = 13.77, P < 0.001; Fig. 1). 3.2. Population structure and latitudinal distribution of wintering individuals 3.2.1. Sex ratio The sex ratio was on average male-biased with a 59% proportion of males (logit-scale intercept estimate = 0.376, SE = 0.050, z = 7.68, P < 0.001). The best model included the covariate ‘Year.c’ (Supplementary Material Table S3), which showed a trend in time toward more males among the residents (logit-scale annual change estimate = 0.011, SE = 0.004, z = 3.07, P = 0.002) (Table 2). There was also a pattern within a typical winter, with an increase in the number of males towards new year, dropping towards more females in the spring (Fig. 2a). The model connecting sex ratio to MRR (study part 1) was not supported. Table 2. The estimated fixed effects coefficients from the best candidate models for sex, age and latitude, and the standard deviations (SD) of the annual random effects. The within-year smoothing functions of these models were all statistically significant and are presented in Fig. 2. “s” before parenthesis denotes a smoothing function. Variable Estimate SE z or t n P Sex, model 1 Intercept 0.376 0.05 7.68 17152 <0.001 Year.c 0.011 0.004 3.07 17152 0.002 s(Dow) 0.025 0.073 0.34 17152 0.733 Rnd SD (Year.f) 0.177 – – 17152 – Age, model 0 Intercept 1.729 0.06 29 5597 <0.001 s(Dow) 0.164 0.037 4.42 5597 <0.001 Rnd SD (Year.f) 0.175 – – 5597 – Latitude, model 3 Intercept 58.85 0.056 1041.8 17152 <0.001 SexMale 0.162 0.033 4.93 17152 <0.001 s(Dow) –0.21 0.217 –0.97 17152 <0.001 Rnd SD (Year.f) 0.262 – – – – Fig. 1. Long-term trend in the annual migrating to wintering ratio of the Eurasian Sparrowhawk (Accipiter nisus). The illustrated regression line and its 95% confidence interval (enclosed by dashed lines), represent the partial effect of (centralized) year from a log-linear multiple regression model. The Y-axis is represented on log-scale. 123 ORNIS FENNICA Vol.101, 2024 3.2.2. Age ratio On average 85% of the reported overwintering sparrowhawks were adults (logit-scale intercept estimate = 1.729, SE = 0.060, z = 29.0, P < 0.001). The best candidate model was a model with no covariates included (Supplementary Material Table S3). There was a within-winter seasonal pattern (Table 2), where the age ratio becomes increasingly dominated by adults towards early spring (Fig. 2b). The model connecting age ratio to MRR was not supported. 3.2.3. Latitudinal distribution The best candidate model was the model with sex included (Supplementary Material Table S3). During winter, males were on average found at higher latitudes compared to females (estimate: 0.162, SE = 0.033, t = 4.93, P < 0.001) (Table 2). Within the winter season, the population weight of residents was found at increasing latitudes, before skewing southward again in early spring (Fig. 2c). The models connecting latitude to MRR were not supported. 4. Discussion 4.1. The migrant-to-resident ratio Deciding whether to migrate or remain in the breeding areas involves balancing the costs and benefits of each strategy, but the ease with which populations and individuals can switch between these behaviours remains understudied. Additionally, the drivers that regulate this phe- notypic variation are not fully understood. On a shorter inter-annual scale, we observed a positive partial effect of late-autumn temperatures on the migrant-to-resident ratio (‘MRR’). This indicates that prevailing temperatures over a small temporal window partly influence the decision to migrate. Moreover, we identified a positive partial trend over the 40-year study period. This response in migration propensity is in a unexpected direction and challenges the hypothesis proposed by Berthold (1996), which suggested that warmer temperatures would reduce the pressure to migrate, thereby lacking a straightforward thermal explanation. Other examples of increased migratoriness include long-term trends in Green- finches (Chloris chloris) in Finland (Meller et al. 2016) and Blue Tits (Cyanistes caeruleus), in Sweden (Nilsson et al. 2006). Berthold’s (1996) hypothesis assumes that all other factors remain Fig. 2. Within-season effects in sex ratio, age ratio and latitudinal location of overwintering Eurasian Sparrowhawks (Accipiter nisus) in Sweden during 1975–2016 (all presented smoothing functions are statistically significant). (a) Males dominated the winter population of residents (horizontal line marks equal probability of male and female) and the probability of a resident being male peaked around new year (December 31 is day 0), while (b) the probability of being adult increased throughout the winter season. (c) The average latitude of males peaked after New Year and gradually became more southerly towards spring. These results are compared to the presented hypotheses of early arrival, body size and competitive release (i.e. dominance). Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 124 relatively unchanged, but this is unlikely over the long term and hence subject to more complex patterns. For example, population trends can lead to density-dependent effects (Kokko & Lundberg 2001, Lundberg 1988), although the sparrowhawk population in Sweden has remained relatively stable during our study period (Green et al. 2014). However, in sparrowhawks, the response to annual temperature likely reflects migration intensity alone (Supplementary Material Table S4), suggesting a facultative decision for at least some individuals in early autumn. Yet, the long-term patterns revealed a complementary increase in migrants and a decrease in residents. In general, food abundance should affect the winter survival of residents and this has earlier been connected with migratoriness in certain terrestrial species (Nilsson et al. 2006, Meller et al. 2016). We found a connection between annual migratoriness and the winter abundance of key prey species. This correlation reflects annual anomalies since the joint abundance of the 10 selected prey species (Supplementary Materials Table S2) showed no overall temporal trend across 40 years. While sparrowhawks gain from being prey specialists during the breeding season (Otterbeck et al. 2015), their dietary niche likely widens and becomes more opportunistic during the winter. The selected prey species may affect the sparrowhawk unequally, and our abundance estimates put more weight on common species, implying that the sparrowhawk responds to wider availability of prey items (in number). The abundance might differ between forest habitats and urban areas—where the latter may facilitate residency (Partecke & Gwinner 2007, Møller et al. 2014; but see Deshpande et al. 2022). We did not account for such habitat differences, but urban areas may play a particular role during harsh winters and also aggregate prey species around bird feeders (Robb et al. 2008a, 2008b; but see Shütz & Schulze 2018). Recent literature has provided cases both where the resident (Grist et al. 2017, Buchan et al. 2020) and the migratory fraction of indi- viduals (Zúñiga et al. 2017, Acker et al. 2021) are suggested to experience a consistent fitness advantage, illustrating how such conditional selection pressures may vary between species, in space and time. There are examples where migratoriness has increased (or even reappeared) in partially migratory species facing marginal en- vironmental conditions. As one, individuals from a resident population of House Finch (Carpodacus mexicanus) were relocated to a colder climate, followed by a reappearance of migratoriness within a few generations (Able & Belthoff 1998). Another more naturally occurring example is the northward breeding range shifts of Serin through Europe, eventually creating a pattern of obligate residents in the south, partial migrants in the middle, and obligate migrants in the north (Mayr 1926, Newton 2008). As the climate subsequently warmed in their new northernmost range, many populations switched back to being partially migratory (Bauer & Berthold 1997). Likewise, within Fennoscandia, the European Robin (Erithacus rubecula) shows an increasing propor- tion of migrants from temperate Denmark in the southwest towards the more continental Finland in the northeast, also coinciding with increased migration distances (Newton 2008). The distribu- tion range of sparrowhawk did expand northwards in Sweden within our study period (Ottvall et al. 2008) gradually towards harsher climate zones (but within the study area). This is consistent with the general pattern seen among birds in response to climate change (Virkkala & Lehikoinen 2014, Välimäki et al. 2016) facilitated by the emergence of areas that were earlier outside of the thermal niche of more southern species. There may still be limits to how well they cope with colder surround- ings relative to northern species (Pakanen et al. 2016) and thus for year-round residency towards higher latitudes. Higher latitudes comprise harsher winters and may thus pose a challenge to southern species (e.g. Pakanen et al. 2016) and affect the overall pressure to migrate (Newton & Dale 1996, Newton 2008, Somveille et al. 2013, Ambrosini et al. 2016) facilitating increased migratoriness within progressively northern populations. While northerly areas may be suitable for breeding, the winters may entail hostile environmental condi- tions forcing local breeders to migrate. Northern areas also have fewer urban habitats, which otherwise might to some degree buffer some of the costs of residency during cold, dark and snowy winters. If true, the southern, middle and northern populations may experience opposing 125 ORNIS FENNICA Vol.101, 2024 selection pressures for migratoriness within larger study areas such as in our study. To validate this as a mechanism, future studies could explore how breeding range shifts affect the average local migratory propensity, ideally across multiple species. Nevertheless, a multi-species approach would likely face the challenge of accounting for a mosaic of species-specific selection pressures which may depend on their morphology, demog- raphy, distribution and life history. The increasing migration numbers at Falsterbo could also partly reflect non-detected long-term population increases (and annual fluc- tuations) in not only Sweden, but also adjacent areas such as Norway and Finland. However, the reported breeding population in Finland has been suggested to decline (Meller et al. 2016) while long-term monitoring lacks from Norway overall. The wintering population in Finland has remained stable compared to the negative Swedish trend in this study, despite that southern Sweden typically provides milder and less con- tinental winter conditions. The opposite trends in migration numbers (increasing) and wintering numbers (decreasing), however, reinforce the result, suggesting that our observed pattern is not solely a result of changes in the breeding population size, nor a methodological artifact in either dataset. 4.2. Demography of migrants and residents The fitness prospects as a resident or a migrant are likely conditional based on individual capabilities, otherwise the one superior strategy would be fixed in the population (Pulido & Berthold 2010; cf. Pulido 2011). In this study, males and adults were over-represented among resident individuals suggesting residency to be a competitive temporal advantage for territory- establishing males in the early spring (Silverin et al. 1989, Grayson & Wilbir 2009, Fudickar et al. 2013, Lehikoinen et al. 2011), while early egg laying increases the clutch size and nestling survival in the species (Otterbeck et al. 2019). By studying a species with reversed sexual size dimorphism, we resolved the problem of overlapping predictions from the traits of being male, establishing territory, and having superior body size. Thus, larger body size seems not an overall decisive trait for residency mid-winter (Fudickar et al. 2013; but cf. Gow & Wiebe 2014, Macdonald et al. 2015). It is a relatively common pattern among partial migrants (Newton 2008), including sparrowhawks (Kjéllen 1992), that juveniles are more prone to migrate than adults which implicitly means that it is common to switch between migratory and resident phenotype at least once. During our study period, the proportion of males among the residents also showed a long-term increase, suggesting that the pressure for early arrival has increased over time for males, decreased for females, or possibly both. Early arrival has also gained support in the past, but mainly through studies focussing on intra-specific migration distances in populations where all individuals migrate, i.e., differential migration (e.g. Cristol et al. 1999, Macdonald et al. 2016). Overwintering may pose an array of chal- lenges and the gain in breeding success needed to outweigh the higher survival costs can be high (Zúñiga et al. 2017). While the subsequent fitness gain by early arrival should outweigh the costs of overwintering to be a superior strategy to year-round residency, the inferior-sized male sparrowhawk may face significant challenges posed by winter conditions. However, we did not find support for a model linking the sex ratio to the annual variation in the proportion of migrants, as these changes were better explained by a mere long-term trend. Within a typical winter, the proportion of males peaked around new year, followed by a drop. It is likely that the left side of this curve partly reflects late female migration in early winter while the decreasing male bias during late winter could even reflect unequal mortality between males and females. Another possibility is an early spring migration of females, which we can not rule out affects the patterns. A more likely factor is that resident males are gradually less observed in agricultural and urban areas as they move closer to their forest breeding territories upon spring. The latter also supports the importance of early arrival due to territory establishment. While we propose that males gain the most from residency, females were present at all latitudes suggesting that some females benefit Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 126 from residency as their best option. Early in the breeding season, for instance, conspecific competition among females may be strong as the number of resident males with an acquired territory (a central resource for females) is initially scarcer than the number of territories (a central resource for males) (Kokko et al. 2006). Another possibility is that the larger-sized females may have relatively low thermal costs by winter temperatures reducing the pressure to migrate, which means that body size may also have an important role for residency among females and should not be disregarded. It remains a possible condition-dependent factor among resident males as well, as having a larger body size (Hegemann et al. 2015) or superior condition (Kokko 1999) should improve the survival prospects of overwintering. It is a common pattern among partial migrants (Newton 2008), including sparrowhawks (Kjéllen 1992), that juveniles are more prone to migrate than adults, which implicitly means that it is common to switch between migratory and resident phenotype at least once. The resident fraction of sparrowhawks consisted of more adults than juveniles, which coincides with juveniles dominating during autumn migration at Falsterbo (e.g. Kjellén 1992, 2019) in line with the part expectations of age in competitive release theory (i.e. juveniles). Yet, juveniles regularly do overwinter, and given the species often breed in the second calendar year, this may benefit territory availability in spring. While there were no long-term trends in the adult/juvenile ratio across the study period, adults became increasing- ly overrepresented as a typical winter progressed. Harsh winter conditions may particularly affect younger and inexperienced individuals, so this could reflect higher mortality of young compared to adults that overwinter.. However, resident males also fit the pattern among juvenile migrants across Falsterbo, where females are overrep- resented among the migrating juveniles; while males show a slight overrepresentation among the smaller fraction of migrating adults (Kjellén 1992). In the neighboring country Finland, which has a more continental climate, adults are even more overrepresented in the autumn migration counts compared to Sweden (Lehikoinen et al. 2014). We found clear latitudinal sex segregation during the winter season (Table 2), with the average latitude of resident males being further north than that of females. The overall (i.e. both sexes) latitudinal distribution span gradually shifted northwards as the first part of the winter season progressed, possibly reflecting early and mid-winter avoidance of the environmental con- ditions prevailing at higher latitudes. It therefore seems plausible that some residents migrate but on small latitudinal scales, combining the better of two worlds, move away from most acute winter conditions while remaining positioned for early spring arrival. This result could also appear if the southernmost individuals migrate (disappear from the country) as the winter proceeds. By not passing any major migratory funnels, however, these movements would seldom be counted by bird observatories such as Falsterbo—the primary data source for studies on bird migration. However, latitudinal patterns could potentially also, at least partly, reflect annual latitudinal patterns in overall observer activity if such exists. 5. Conclusions We studied partial migration from the viewpoint of the small raptor, the Eurasian Sparrowhawk (Accipiter nisus), through two different but interconnected perspectives: 1) how the migrant to resident ratio changed over time, responded to short-term autumn temperatures, and food availa- bility, 2) what underlying individual demograph- ic traits may affect the expression of a migratory or a resident phenotype. There was a strong long-term temporal trend towards increased mi- gratoriness across the study period. The migrant to resident ratio unexpectedly increased with higher late-autumn temperatures, and expectedly decreased with higher winter food abundance. The average winter sex ratio of residents was male-biased, which supports the arrival time hypothesis. The sex ratio also became increas- ingly male-dominated across the study period. However, the vast majority of the overwintering individuals were adults, which fulfils part as- sumptions from competitive release hypothesis, but showing no trends. The proportion of adults increased throughout the winter, which may 127 ORNIS FENNICA Vol.101, 2024 reflect asymmetric mortality. The distribution of males was typically more northerly stretched than females and moved gradually towards higher latitudes as the winter progressed. This suggests some form of small-scale migration among individuals typically considered residents. We present a picture of migration being a conditional strategy among partial migrants while small-scale migratory patterns also occur adaptively among residents. Osittain muuttavan varpushaukan (Accipiter nisus) muuttoalttius ja paikalle jäävän osuuden koirasvoittoisuus kasvavat Ruotsissa Osittaismuuttajien populaatiot koostuvat sekä muuttavista että paikallisista yksilöistä. Muut- tavat ja paikalliset kohtaavat talviaikaan erilaisia ekologisia olosuhteita, ja niiden muutto- päätökseen vaikuttavat taustatekijät ovat edelleen kiistanalaisia sekä populaation että yksilön näkö- kulmasta. Tässä tutkimuksessa tarkastelimme varpushaukan muuttoliikehdintää kahdesta eri, mutta toisiinsa kytkeytyvästä näkökulmasta: 1) selvitimme muttoalttiudessa (muuttajien määrissä suhteessa paikallisten määriin) esiin- tyviä trendejä ja vaihtelua populaatiotasolla ja 2) tarkastelimme, onko ikä ja sukupuoli yhtey- dessä yksilön alttiuteen muuttaa tai talvehtia. Aineistona käytimme kansalaishavaintoaineistoa neljän vuosikymmenen ajalta, jonka avulla ana- lysoimme talvehtivien varpushaukkojen ikä- ja sukupuolijakauman ajallista ja alueellista vaihte- lua Ruotsissa. Havaitsimme odotustemme vastaisesti, että vuosittainen muuttoalttius kasvoi ajan myötä ja oli suurempi mitä lämpimämpi syksy oli kyseessä. Lisäksi muuttavien varpushaukko- jen osuus kasvoi, kun talvisen saaliseläinten määrä oli pienempi. Keskimääräinen talven sukupuolijakauma oli koiraspainotteinen, ja tämä vinouma kasvoi vuosien varrella. Esitäm- mekin, että talvehtiminen hyödyttää reviirejä hakevia koiraita, sillä aikainen läsnäolo antaa kilpailuedun korkealaatuisten reviirien hankin- nassa. Lisäksi Ruotsissa talvehtivien yksilöiden levinneisyys (sukupuolesta riippumatta) siirtyi asteittain pohjoisemmaksi talven edetessä, mikä viittaa pienimuotoiseen muuttoliikkeeseen pai- kallisen populaation keskuudessa. Nämä tulokset tarjoavat uusia näkökulmia osittaismuuton taus- tatekijöihin ja säätelyyn. Acknowledgements. We want to thank two anonymous referees and Anssi Vähätalo for their valuable comments on the manuscript. A big thanks to Aleksi Lehikoinen for helpful input during the process. We are grateful for the many volunteers who contributed to the Swedish Bird Survey. Andreas Otterbeck was funded by Societas pro Fauna et Flora Fennica and Novia University of Applied Sciences. We want to thank Lennart Carlsson for help with extracting the Swedish dataset from the Swedish ‘Artportalen’. References Able, K.P. & Belthoff, J.R. 1998: Rapid ‘evolution’ of migratory behaviour in the introduced house finch of eastern North America. — Proceedings of the Royal Society of London B: Biological Sciences 265: 2063– 2071. https://doi.org/10.1098/rspb.1998.0541 Acker, P., Daunt, F., Wanless, S., Burthe, S.J., Newell, M.A., Harris, M.P., Grist, H., Sturgeon, J., Swann, R.L., Gunn, C., Payo-Payo, A. & Reid, J.M. 2021: Strong survival selection on seasonal migration versus residence induced by extreme climatic events. — Journal of Animal Ecology 90: 796–808. https://doi. org/10.1111/1365-2656.13410 Ambrosini, R., Cuervo, J.J., du Feu, C., Fiedler, W., Musitelli, F., Rubolini, D., Sicurella, B., Spina, F., Saino, N. & Møller, A.P. 2016: Migratory connectivity and effects of winter temperatures on migratory behaviour of the European robin Erithacus rubecula: a continent–wide analysis. — Journal of Animal Ecology 85: 749–760. https://doi.org/10.1111/1365-2656.12497 Arnold, T.W. 2010: Uninformative Parameters and Model Selection Using Akaike’s Information Criterion. — The Journal of Wildlife Management 74: 1175–1178. https:// doi.org/10.1111/j.1937-2817.2010.tb01236.x Bauer, H.G. & Berthold, P. 1997: Die Brutvogel Mittel- europas. Bestand und Gefährdung. — Aula-Verlag, Wiesbaden. (In German) Belthoff, J.R. & Gauthreaux, S.A. 1991: Partial Migration and Differential Winter Distribution of House Finches in the Eastern United States. — The Condor 93: 374– 382. https://doi.org/10.2307/1368953 Berthold, P. 1988: Evolutionary aspects of migratory behavior in European warblers. — Journal of Evolutionary Biology 1: 195–209. https://doi. org/10.1046/j.1420-9101.1998.1030195.x Berthold, P. 1996: Control of Bird Migration. — The Auk 114(3): 534–535. https://doi.org/10.2307/4089262 Berthold, P. 1999. A comprehensive theory for the evolution, control and adaptability of avian migration. — Ostrich Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 128 70: 1–11. https://doi.org/10.1080/00306525.1999.9639 744 Berthold, P. 2001: Bird migration: a general survey. — Oxford University Press, Oxford Berthold, P. & Querner, U. 1982: Genetic basis of moult, wing length, and body weight in a migratory bird species, Sylvia atricapilla. — Experientia 38: 801–802. https://doi.org/10.1007/BF01972279 Biebach, H. 1983: Genetic determination of partial migration in the European Robin (Erithacus rubecula). — The Auk 100: 601–606. https://doi.org/10.1093/ auk/100.3.601 Boyle, W.A., Norris, D. R. & Guglielmo, C. G. 2010: Storms drive altitudinal migration in a tropical bird. — Proceedings of the Royal Society of London B: Biological Sciences 277: 2511–2519. https://doi. org/10.1098/rspb.2010.0344 Buchan, C., Gilroy, J.J., Catry, I. & Franco, A.M.A. 2020: Fitness consequences of different migratory strategies in partially migratory populations: a multi-taxa meta- analysis. — Journal of Animal Ecology 89: 678–690. https://doi.org/10.1111/1365-2656.13155 Chapman, B.B., Brönmark, C., Nillson, J.Å. & Hansson, L.A. 2011: The ecology and evolution of partial migration. — Oikos 120: 1764–1775. https://doi. org/10.1111/j.1600-0706.2011.20131.x Cristol, D.A., Baker, M.B & Carbone, C. 1999: Differential Migration Revisited. — In Current Ornithology vol 15 (ed. Nolan, V., Ketterson, E.D. & Thompson, C.F.). Springer, Boston, MA. https://doi.org/10.1007/978-1- 4757-4901-4_2 Daan, S., Masman, D. & Groenewold, A. 1990: Avian basal metabolic rates: their association with body composition and energy expenditure in nature. — The American Journal of Physiology 259: 333–340. https://doi. org/10.1152/ajpregu.1990.259.2.r333 Deshpande, P., Lehikoinen, P., Thorogood, R. & Lehikoinen, A. 2022: Snow depth drives habitat selection by overwintering birds in built-up areas, farmlands and forests. — Journal of Biogeography 49: 630–639. https://doi.org/10.1111/jbi.14326 de Zoeten, T. & Pulido, F. 2020: How migratory populations become resident. — Proceedings of the Royal Society of London B: Biological Sciences B 287: 20193011. https://doi.org/10.1098/rspb.2019.3011 Dunning Jr., J.B. (ed) 2008: CRC handbook of avian body masses. — CRC press, London. https://doi. org/10.1201/9781420064452 Fudickar, A.M., Schmidt, A., Hau, M., Quetting, M. & Partecke, J. 2013: Female‐biased obligate strategies in a partially migratory population. — Journal of Animal Ecology 82: 863–871. https://doi.org/10.1111/1365- 2656.12052 Gauthreaux, S.A. 1982: The ecology and evolution of avian migration systems. — Avian Biology 6: 93–168. Gow, E.A. & Wiebe, K.L. 2014: Males migrate farther than females in a differential migrant: an examination of the fasting endurance hypothesis. — Royal Society Open Science 1: 140346. https://doi.org/10.1098/rsos.140346 Grayson, K.L. & Wilbur, H.M. 2009: Sex- and context- dependent migration in a pond-breeding amphibian. — Ecology 90: 306–311. Green, M. & Lindström, Å. 2014: Övervakning av fåglarnas populationsutveckling. Årsrapport för 2013. — Institute of Biology, University of Lund. (In Swedish) Grist, H., Daunt, F., Wanless, S., Burthe, S.J., Newell, M.A., Harris, M.P. & Reid, J.M. 2017: Reproductive performance of resident and migrant males, females and pairs in a partially migratory bird. — Journal of Animal Ecology 86: 1010–1021. https://doi.org/10.1111/1365- 2656.12691 Hegemann, A., Marra, P.P. & Tieleman, B.I. 2015: Causes and consequences of partial migration in a passerine bird. — The American Naturalist 186: 531–546. https:// doi.org/10.1086/682667 Karlsson, L. 2004: Wings over Falsterbo. — Falsterbo Bird Observatory, Falsterbo. Ketterson, E.D. & Nolan Jr, V. 1976: Geographic variation and its climatic correlates in the sex ratio of eastern–wintering Dark–eyed Juncos (Junco hyemalis hyemalis). — Ecology 57: 679–693. https://doi. org/10.2307/1936182 Kjellén, N. 1992: Differential timing of autumn migration between sex and age groups in raptors at Falsterbo, Sweden. — Ornis Scandinavica 23: 420–434. https:// doi.org/10.2307/3676673 Kjellén, N. 2019: Migration counts at Falsterbo, Sweden. — Bird Census News 32: 27–37. Retrieved from https://www.ebcc.info/wp-content/uploads/2020/06/4- kjellen-32-1-2.pdf Kokko, H. 1999: Competition for early arrival in migratory birds. — Journal of Animal Ecology 68: 940–950. https://doi.org/10.1046/j.1365-2656.1999.00343.x Kokko, H. & Lundberg, P. 2001: Dispersal, migration, and offspring retention in saturated habitats. — The American Naturalist 157: 188–202. http://dx.doi. org/10.1086/318632 Kokko, H., Gunnarson, T.G., Morrell, L.J. & Gill, J.A. 2006: Why do female migratory birds arrive later than males? — Journal of Animal Ecology 75: 1293–1303. https:// doi.org/10.1111/j.1365-2656.2006.01151.x Lack, D. 1943: The problem of partial migration. — British Birds 37: 122–130. Lack, D. 1944: The problem of partial migration. — British Birds 37: 143–150. Lehikoinen, A. 2011: Advanced autumn migration of sparrowhawk has increased the Predation risk of long- distance migrants in Finland. — PLoS ONE 6: e20001. https://doi.org/10.1371/journal.pone.0020001 Lehikoinen, A., Hokkanen, T. & Lokki, H. 2011: Young and female-biased irruptions in pygmy owls Glaucidium passerinum in southern Finland. — Journal of Avian Biology 42: 564–569. https://doi.org/10.1111/ j.1600-048X.2011.05461.x 129 ORNIS FENNICA Vol.101, 2024 Lehikoinen, A., Ekroos, J., Piha, M., Seimola, T., Tirri, I.S, Velmala, W. & Vähätalo, A. 2014: Muuton ajoittuminen eri ikäluokilla ja sukupuolilla Hangon lintuasemalla rengastuksen perusteella: Osa 1: syksyiset ei- varpuslinnut. — Tringa 41: 30–53. (In Finnish with English summary) Lehikoinen, A., Lindén, A., Karlsson, M., Andersson, A., Crewe, T.L., Dunn, ... & Tjørnløv, R. S. 2019: Phenology of the avian spring migratory passage in Europe and North America: asymmetric advancement in time and increase in duration. — Ecological Indicators 101: 985– 991. https://doi.org/10.1016/j.ecolind.2019.01.083 Lindén, A., Lehikoinen, A., Hokkanen, T. & Väisänen, R.A. 2011: Modelling irruptions and population dynamics of the great spotted woodpecker–joint effects of density and cone crops. — Oikos 120: 1065–1075. http://dx.doi. org/10.1111/j.1600-0706.2010.18970.x Lundberg, P. 1985: Dominance behaviour, body weight and fat variations, and partial migration in European blackbirds Turdus merula. — Behavioral Ecology and Sociobiology 17: 185–189. https://doi.org/10.1007/ BF00299250 Lundberg, P. 1988: The evolution of partial migration in birds. — Trends in Ecology & Evolution 3: 172–175. https://doi.org/10.1016/0169-5347(88)90035-3 Lundblad, C.G. & Conway, C.J. 2020: Testing four hypotheses to explain partial migration: balancing reproductive benefits with limits to fasting endurance. — Behavioral Ecology and Sociobiology 74: 1–16. https://doi.org/10.1007/s00265-019-2796-3 Macdonald, C.A., McKinnon, E.A., Gilchrist, H.G. & Love, O.P. 2016: Cold-tolerance, and not earlier arrival on breeding grounds, explains why males winter further north in an Arctic-breeding songbird. — Journal of Avian Biology 47: 7–15. https://doi.org/10.1111/ jav.00689 Main, I. G. 2002: Seasonal movements of Fennoscandian Blackbirds Turdus merula. — Ringing & Migration 21: 65–74. Mayr, E. 1926: Die ausbreitung des Girlitz (Serinus canaria serinus L.). — Journal für Ornithologie 74: 571–671. (In German) Meller, K., Vähätalo, A.V., Hokkanen, T., Rintala, J., Piha, M. & Lehikoinen, A. 2016: Interannual variation and long–term trends in proportions of resident individuals in partially migratory birds. — Journal of Animal Ecology 85: 570–580. https://doi.org/10.1111/1365- 2656.12486 Møller, A.P., Jokimäki, J., Skorka, P. & Tryjanowski, P. 2014: Loss of migration and urbanization in birds: a case study of the blackbird (Turdus merula). — Oecologia 174: 1019–1027. https://doi.org/10.1007/ s00442-014-2953-3 Newton, I. 1986: The Sparrowhawk. — T & Poyser, Calton. Newton, I. 2008: The migration ecology of birds. — Academic Press, Elsevier, London. Newton, I. & Dale, L. 1996: Relationship between migration and latitude among west European birds. — Journal of Animal Ecology 65: 137–146. https://doi. org/10.2307/5716 Nilsson, A.L., Lindstroem, A., Jonzén, N., Nilsson, S.G. & Karlsson, L. 2006: The effect of climate change on partial migration–the blue tit paradox. — Global Change Biology 12: 2014–2022. https://doi.org/10.1111/ j.1365-2486.2006.01237.x Otterbeck, A., Lindén, A. & Roualet, É. 2015: Advantage of specialism: reproductive output is related to prey choice in a small raptor. — Oecologia 179: 129–137. http:// dx.doi.org/10.1007/s00442-015-3320-8 Otterbeck, A., Selås, V., Nielsen, J. T., Roualét, E. & Lindén, A. 2019: The paradox of nest reuse: early breeding benefits reproduction, but nest reuse increases nest predation risk. — Oecologia 190: 559–568. https://doi. org/10.1007/s00442-019-04436-7 Ottvall, R., Edenius, L., Elmberg, J., Engström, H., Green, M., Holmqvist, N., Lindström, Å., Tjernberg, M. & Pärt, T. 2008: Populationstrender för fågelarter som häckar i Sverige. Report 5813. — Naturvårdsverket, Stockholm. (In Swedish) Pakanen, V.M., Ahonen, E., Hohtola, E. & Rytkönen, S. 2018: Northward expanding resident species benefit from warming winters through increased foraging rates and predator vigilance. — Oecologia 188: 991–999. https://doi.org/10.1007/s00442-018-4271-7 Pannekoek, J. & Van Strien, A. 2005: TRIM 3 manual (Trends &indices for monitoring data). — JM Voorburg, Statistics Netherlands, The Netherlands. Partecke, J. & Gwinner, E. 2007: Increased sedentariness in European Blackbird following urbanization: a consequence of local adaptation? — Ecology 88: 882– 890. https://doi.org/10.1890/06-1105 Pulido, F. 2011: Evolutionary genetics of partial migration – the threshold model of migration revis(it)ed. — Oikos 120: 1776–1783. https://doi. org/10.1111/j.1600-0706.2011.19844.x Pulido, F. & Berthold, P. 2010: Current selection for lower migratory activity will drive the evolution of residency in a migratory bird population. — Proceedings of the National Academy of Sciences of the United States of America 107: 7341– 7346. https://doi.org/10.1073/ pnas.0910361107 R Core Team 2018: R: A language and environment for statistical computing. — R Foundation for Statistical Computing, Vienna, Austria. Risch, M. & Brinkhof, M. W. G. 2002: Sex ratios of Sparrowhawk (Accipiter nisus) broods: the importance of age in males. — Ornis Fennica 79: 49–59. Robb, G.N., Mcdonald, R.A., Chamberlain, D.E. & Bearhop, S. 2008a: Food for thought: supplementary feeding as a driver of ecological change in avian populations. — Frontiers in Ecology and the Environment 6: 476–484. http://dx.doi.org/10.1890/060152 Robb, G.N., Mcdonald, R.A., Chamberlain, D.E., Reynolds, S.J., Harrison, T.J. & Bearhop, S. 2008b: Winter feeding Otterbeck & Lindén: Shifts in migration and male bias in Swedish sparrowhawks 130 of birds increases productivity in the subsequent breeding season. — Biology Letters 4: 220–223. https:// doi.org/10.1098/rsbl.2007.0622 Rubolini, D., Møller, A.P., Rainio, K. & Lehikoinen, E. 2007: Intraspecific consistency and geographic variability in temporal trends of spring migration phenology among European bird species. — Climate Research 35: 135–146. http://dx.doi.org/10.3354/ cr00720 Schütz, C. & Schulze, C.H. 2018: Park size and prey density limit occurrence of Eurasian Sparrowhawks in urban parks during winter. — Avian Research 9: 30. https:// doi.org/10.1186/s40657-018-0122-9 Schwabl, H. 1983: Ausprägung und Bedeutung des Teilzugverhaltens einer südwestdeutschen Population der Amsel Turdus merula. — Journal für Ornithologie 124: 101–116. (In German) Silverin, B., Viebke, P.A. & Westin, J. 1989: Hormonal correlates of migration and territorial behavior in juvenile willow tits during autumn. — General and Comparative Endocrinology 75(1): 148–156. https:// doi.org/10.1016/0016-6480(89)90020-8 Smith, H.G. & Nilsson, J.Å. 1987: Intraspecific variation in migratory pattern of a partial migrant, the blue tit (Parus caeruleus): an evaluation of different hypotheses. — The Auk 104: 109–115. https://doi.org/10.2307/4087239 Somveille, M., Manica, A., Butchart, S.H.M. & Rodrigues, A.S.L. 2013: Mapping global diversity patterns for migratory birds. — PLoS ONE 8: e70907. https://doi. org/10.1371/journal.pone.0070907 Terrill, S.B. & Able, K.P. 1988: Bird Migration Terminology. — The Auk 105: 205–206. https://doi.org/10.1093/ auk/105.1.205 Usui, T., Butchart, S.H.M. & Phillimore, A.B. 2017: Temporal shifts and temperature sensitivity of avian spring migratory phenology: a phylogenetic meta- analysis. — Journal of Avian Ecology 86: 250–261. https://doi.org/10.1111/1365-2656.12612 Van Vliet, J., Musters, C.J.M. & Ter Keurs, W.J. 2009: Changes in migration behaviour of Blackbirds Turdus merula from the Netherlands. — Bird Study 56: 276– 281. https://doi.org/10.1080/00063650902792148 Välimäki, K., Lindén, A. & Lehikoinen, A. 2016: Velocity of density shifts in Finnish land bird species depends on their migration ecology and body mass. — Oecologia 181: 313–321. https://doi.org/10.1007/s00442-015- 3525-x Visser, M.E. & Gienapp, P. 2019: Evolutionary and demographic consequences of phenological mismatches. — Nature Ecology & Evolution 3: 879– 885. https://doi.org/10.1038/s41559-019-0880-8 Virkkala, R. & Lehikoinen, A. 2014: North by north-west: climate change and directions of density shifts in birds. — Global Change Biology 22: 1121–1129. https://doi. org/10.1111/gcb.13150 Wood, S.N. 2017: Generalized Additive Models: An Introduction with R (2nd edition). — Chapman and Hall/CRC, New York. https://doi. org/10.1201/9781315370279 Wood, S. & Scheipl, F. 2020: gamm4: Generalized Additive Mixed Models using ‘mgcv’ and ‘lme4’. R package version 0.2-6. Accessed at https://CRAN.R-project.org/ package=gamm4/ Zúñiga, D., Gager, Y., Kokko, H., Fudickar, A. M., Schmidt, A., Naef-Daenzer, B., Wikelski, M. & Partecke, J. 2017: Migration confers winter survival benefits in a partially migratory songbird. — eLife 6: e28123. https://doi. org/10.7554/eLife.28123 Online supplementary material Supplementary material available in the online version of the article (https://doi.org/10.51812/ of.122172) includes Figure S1 and Tables S1–S4.