Introduction
Detecting factors causing declines in wildlife populations provides essential knowledge for the development of effective conservation strategies (Salafsky et al. Reference Salafsky, Margoluis, Redford and Robinson2002; Faaborg et al. Reference Faaborg, Holmes, Anders, Bildstein, Dugger and Gauthreaux2010). Variation in demographic parameters, which is frequently age specific, guides the management of wildlife populations (Sandercock Reference Sandercock2006). Among demographic parameters, survival is often thought to have the greatest impact on population dynamics, especially among long-lived species (Doherty et al. Reference Doherty, Schreiber, Nichols, Hines, Link and Schenk2004; Schorcht et al. Reference Schorcht, Bontadina and Schaub2009; Sandercock Reference Sandercock, Murray and Sandercock2020). However, in wild populations, separating survival from dispersal is challenging because it is often impossible to distinguish between the death and emigration of an animal. Although various statistical models have been developed to distinguish between the two parameters, real data are often insufficient for the application of these models (Sandercock Reference Sandercock, Murray and Sandercock2020). Therefore, “apparent survival”, defined as the product of real survival and site fidelity, is often estimated (Lebreton et al. Reference Lebreton, Burnham, Clobert and Anderson1992).
The centre–periphery hypothesis assumes that environmental conditions are more favourable for populations located near the centre of the range and harsher at the periphery (Brown Reference Brown1984). Both biotic and abiotic factors determine species distributions (Brown et al. Reference Brown, George and Kaufman1996), and in parallel, influence population size and demographic processes across the range (Pironon et al. Reference Pironon, Papug, Villellas, Angert, García and Tompson2016). For example, bird species tend to decrease in abundance from the centre to the periphery (Väli et al. Reference Väli, Treinys and Lõhmus2004; Pironon et al. Reference Pironon, Papug, Villellas, Angert, García and Tompson2016) and show lower demographic rates in populations located at the margins of distributions (Niedzielski and Bowman Reference Niedzielski and Bowman2014; Konovalov et al. Reference Konovalov, Nellis, Nellis, Nurmla, Sellis and Väli2019), although several studies have failed to detect spatial trend in demographic rates (e.g. Treinys et al. Reference Treinys, Väli and Berganis2017; Martín et al. Reference Martín, Onrubia and Ferrer2021). These differences in demographic properties along the centre–periphery axis emphasise the importance of spatial analyses of key demographic variables, especially for threatened species or species containing regionally threatened populations.
Black Stork Ciconia nigra is a large, long-lived bird species with a breeding distribution across Europe and Asia and in southern Africa. In Europe, it has an unfavourable conservation status and is listed in Annex I of the EU Birds Directive (2009/147/EC). This migratory species is also listed in Annex II of the Berne, Bonn, and Convention on International Trade in Endangered Species of Wild fauna and Flora (CITES) Conventions, and is included in the Agreement on the Conservation of African-Eurasian Migratory Waterbirds (AEWA). The Black Stork nearly disappeared from western Europe in the first half of the last century; however, populations have been recovering in recent decades (Janssen et al. Reference Janssen, Hormann and Rohde2004; Cano Alonso and Strazds Reference Cano Alonso, Strazds, Keller, Herrando, Voříšek, Franch, Kipson and Milanesi2020; European Commission 2020). The opposite pattern has been observed in northern Europe, where the population was once large but has declined since the 1980s (Janssen et al. Reference Janssen, Hormann and Rohde2004; Cano Alonso and Strazds Reference Cano Alonso, Strazds, Keller, Herrando, Voříšek, Franch, Kipson and Milanesi2020; European Commission 2020). Although previously reported abundances may be somewhat biased, the declining trends are obvious and alarming. For example, the Estonian population decreased by more than three times in 1991–2020 (Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b). In Latvia, the numbers declined by 55% between the early 1990s and early 2000s (Strazds Reference Strazds2011). In Lithuania, the decrease between the 1980s and mid-2000s was at least 20% (Treinys et al. Reference Treinys, Lõhmus, Stončius, Skuja, Drobelis and Šablevičius2008), with a continuous decline thereafter (Treinys Reference Treinys and Rašomavičius2021). A two-fold decline was recorded in north-eastern Poland in 1989–2019 (Zawadzki et al. Reference Zawadzki, Zawadzki, Drozdowski and Zawadzka2022). Owing to severe declines and/or small population sizes, the Black Stork is considered endangered and strictly protected in the above-mentioned countries.
The secretive habits of this woodland-nesting species have hindered studies of its biology (Strazds Reference Strazds2011), and analyses of its survival are limited. Re-sightings of colour-ringed birds have been thoroughly analysed only in Hungary (Central Europe), where apparent survival was low in the first year of life (0.16), but much higher later (0.83; Tamás Reference Tamás2011). A preliminary study in Estonia (northern Europe) reported a two-fold lower estimate for first-year (1y) birds (0.08), but a similar survival rate for older birds (0.84; Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b). Väli et al. (Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b) also compared the two sexes and noticed a remarkably lower apparent survival rate among females. Although the analysis was based on a limited sample, this difference between sexes would explain the high proportion of unpaired birds, low reproductive success, and population decline in Estonia (Konovalov et al. Reference Konovalov, Nellis, Nellis, Nurmla, Sellis and Väli2019). Uncovering the factors contributing to the rapidly decreasing population in northern Europe is an urgent aim requiring more extensive analyses of species demography.
In the current study, we used long-term data from re-sighted colour-ringed birds to estimate age- and sex-specific apparent survival in a declining Black Stork population in a region located east of the Baltic Sea (Estonia, Latvia, and Lithuania, hereafter the Baltic region). Assuming that the demographic patterns of the Black Stork population follow the centre–periphery hypothesis, we expected that (H1) survival rates are lower in the declining Baltic population than in an increasing Hungarian population (European Commission 2020), and also hypothesised that (H2) survival decreases northwards within the Baltic region (northern part of the range) due to the increase in migration distance and severity of environmental conditions. We also expected that (H3) females have a lower survival rate than that of males, similar to most birds (e.g. Donald Reference Donald2007; Newton et al. Reference Newton, McGrady and Oli2016). Finally, we hypothesised that (H4) survival rates in the Baltic Black Stork population decreased over time along with the declining abundance. For most analyses, we used a large sample of colour-ringed birds. In addition, we estimated the survival of satellite-tracked birds, enabling us to exclude the effect of dispersal and providing estimates of real survival.
Methods
Study area and species
The study was conducted in the Baltic region (including Estonia, Latvia, and Lithuania), northern Europe, on the northern limit of the European distribution of the species (Figure 1). Baltic storks are long-distance migrants, covering some 6,000 km to reach their wintering grounds in eastern sub-Saharan Africa and the Middle East; they arrive at breeding grounds in late March or early April and depart in late August or early September (Sellis Reference Sellis2012; Elliott et al. Reference Elliott, Christie, Garcia, Boesman, del Hoyo, Elliott, Sargatal, Christie and de Juana2014, our data). The nests are located in woodlands on large trees, and storks forage primarily on fish in forest streams and other aquatic habitats. Typically, the same nest is used for many years: accordingly, re-sighting efforts can be focused on nest sites, which facilitates research on demographic processes in the breeding population.
Figure 1. Distribution of the Black Stork Ciconia nigra in central and eastern parts of Europe. Crosshatching indicates permanent breeding sites (according to Cano Alonso and Strazds Reference Cano Alonso, Strazds, Keller, Herrando, Voříšek, Franch, Kipson and Milanesi2020), simple hatching indicates scattered or irregular breeding sites. The study area is indicated with bold borders (EE: Estonia, LV: Latvia, LT: Lithuania); HU denotes Hungary, where survival was studied earlier.
Regular monitoring of Black Stork nests in the Baltic region started in Latvia in 1978 (Strazds Reference Strazds2011), followed by Estonia in 1991 (Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b), and Lithuania since the early 2000s (Treinys et al. Reference Treinys, Lõhmus, Stončius, Skuja, Drobelis and Šablevičius2008). Nests are usually inspected from late May until the end of July; however, nestlings are usually ringed in late June or early July when they are at least three weeks old (Strazds et al. Reference Strazds, Ķuze and Reine2006). In the course of monitoring, all nestlings are ringed with scientific (metal) rings. Since 1990 (Latvia), 1994 (Estonia), and 2006 (Lithuania), plastic colour rings with unique alphanumeric codes have been added to the tibiotarsus of the other leg. This facilitates the re-sighting of live individuals without the need for trapping. Over the study period, only three losses of colour rings were recorded. Therefore, similarly to analogous studies of other species (e.g. Wahl and Barbraud Reference Wahl and Barbraud2014; Väli et al. Reference Väli, Kalvāns and Tuvi2021a), these were considered very rare events and our mark–re-sighting analyses did not account for ring losses.
Data collection
Prior to 2010, data for re-sightings were obtained from field observations (mostly photographs). Later, data were largely collected using camera traps (trail cameras) installed at nests in Estonia and Latvia (15–25 in 2010–2015, 30–50 in 2016–2022), and additional data were obtained from five webcams set up at nests. Cameras were mounted on the branch of the nesting tree, or on the neighbouring tree, 1–5 m (usually 2–3 m) from the nest. Cameras were installed in the second half of March, before storks returned from wintering grounds to their nest sites, or during the ringing of nestlings in late June or early July (see further methodological details in Konovalov et al. Reference Konovalov, Nellis, Nellis, Nurmla, Sellis and Väli2019). Few ringed birds were recorded by camera traps installed at foraging sites (streams and rivulets, usually supplied with additional food to attract and trap storks) and one was recorded by a hunter’s camera trap. No camera traps or webcams have been used in Lithuania.
The data set was composed of encounter histories of 3,304 nestlings (539 from Estonia, 1,881 from Latvia, and 825 from Lithuania) marked with colour rings in 1990–2020; re-sightings were collected up to 2022. A total of 91 storks (12 from Estonia, 60 from Latvia, and 19 from Lithuania) were re-sighted at least once in subsequent years alive in the Baltic region. A subsample of 175 birds from Estonia and Latvia were sexed by DNA analysis (Fridolfsson and Ellegren Reference Fridolfsson and Ellegren1999). This subsample (79 females and 96 males), supplemented with marked but unsexed and later unseen birds (assuming an equal sex ratio in the Baltic region; Konovalov et al. Reference Konovalov, Kaldma, Bokotey, Brossault, Chapalain and Dmitrenok2015), was used for detecting sex-specific biases in survival.
Finally, real survival was estimated for birds tagged with GPS transmitters in 2005–2021 (https://birdmap.5dvision.ee/, M. Strazds et al., unpublished data), including 45 juveniles (27 females, 5 males, and 13 of unknown sex; 28 from Latvia, 17 from Estonia) and 26 adults (17 males, 7 females, 2 of unknown sex; 23 from Estonia, 3 from Latvia). Each of these birds was equipped with a 30–50 g (<3% of the body mass) solar-powered GPS tag (Microwave Telemetry Inc., Ecotone Telemetry, Ornitela UAB, E-OBS GmbH), as a backpack, using full Teflon harnesses. Thirty-six GPS transmitters were collected in the field after the suspected death of the bird and in all cases mortality was confirmed; there were no known losses of GPS loggers nor confirmed failures over the years.
We also evaluated natal dispersal as the distance and direction between the nest of hatching (and ringing) and the nest of the first known breeding (only birds ringed as nestlings in Latvia were included in this analysis). Long-lived Black Storks do not breed every year. When storks are not breeding, they may explore several nests in the area. In subsequent years, breeding of the previously exploring individual is recorded in one of the explored nests (both sexes do so, as verified by repeated re-sightings and GPS tracking; unpublished data of authors). Therefore, to increase sample size, we supplemented the verified breeding records with observations at nests even when the bird was not breeding there. Although this approach may cause some bias in the results, the bias is not likely to be large and is equal for both sexes; hence the comparison between sexes is not affected.
Data analysis
Data were analysed using R.4.0.0 (R Core Team 2023). Analyses only included re-sightings in the Baltic region (breeding grounds). Probabilities of re-sighting (encounter; p) and apparent survival (Φ) were estimated using RMark 2.2.7 (Laake Reference Laake2013), an R-package for model building in MARK (White and Burnham Reference White and Burnham1999). The real survival of GPS-tracked birds was also estimated by RMark, with the probability of re-sighting fixed at 1. According to White and Burnham (Reference White and Burnham1999), we initially generated a fully time-dependent Cormack–Jolly–Seber model with “sex” and/or “country” as groups. The goodness-of-fit test (release.gof function in RMark) indicated a poor fit of this model (Test 2: P <0.97, Test 3: P <0.001), suggesting a hidden structure within our sample in which survival patterns differed among groups of birds. This was consistent with earlier results of Tamás (Reference Tamás2011), who distinguished two age classes with differences in survival: (1) birds during their first year of life (1y birds) and (2) older birds. Therefore, we assumed the same structure to test our first hypothesis (H1) that apparent survival is lower in the declining Baltic population than in a previously observed stable population (Tamás Reference Tamás2011). Additionally, we were aware of changes in observation effort over time and incorporated time-dependent re-sighting probability in the model. Thus, our initial model was Φa2 pt. For comparison, we evaluated mean re-sighting probabilities for the two groups using the simplified model Φa2 p2. To evaluate the hypotheses of survival differences between the three Baltic countries (H2) and two sexes (H3), country and sex, respectively, were included in the model (Φa2g pt; Φa2s pt). Finally, we included time dependence to evaluate whether survival decreased over time along with declining abundance (Φt pt; H4). In this analysis, groups were not considered owing to a low sample size. The models were compared using Akaike’s information criterion corrected for small samples (AICc). Following Cooch and White (Reference Cooch and White2016), we evaluated the fit of all models using the Fletcher variance inflation factor ĉ (Fletcher Reference Fletcher2012), which is more appropriate for data sets involving a large number of capture histories with sparse data and low expected re-sighting frequencies than other overdispersion parameters (Cooch and White Reference Cooch and White2016). The models fitted to the data well (Fletcher ĉ = 1.00). To study sex-specific differences in natal dispersal distances and directions, the nonparametric Wilcoxon rank-sum test and Watson–Wheeler test (package Circular; Agostinelli and Lund Reference Agostinelli and Lund2023), respectively, were used.
Results
Apparent survival in the Baltic population (H1; model Φa2 pt) was 0.043 ± 0.005 (SE; 95% confidence interval [CI] = 0.034–0.054) among 1y birds and 0.886 ± 0.016 (0.850–0.914) among older birds. Given the substantial increase in study effort after the mid-2000s, we additionally estimated apparent survival using the subset of data from 2007–2022. The survival estimates for 1y birds (0.042 ± 0.006; 0.032–0.055) and older birds (0.865 ± 0.025; 0.809–0.907) were similar to those estimated from the full data set. The estimated real survival of GPS-tracked 1y birds was also very low (0.11 ± 0.05; 0.05–0.24), yet higher than that of the colour-ringed birds. Real survival of GPS-tracked older (adult) birds from Estonia and Latvia was slightly lower (0.77 ± 0.04; 0.68–0.85) than the estimated apparent survival.
To evaluate the possibility that different re-sighting probabilities account for the low survival estimate among 1y birds, we fitted the data in the simplified model including only two age classes (Φa2 pt). The re-sighting probability was lower among 1y birds (0.09 ± 0.03) than that among older birds (0.23 ± 0.02), while the estimates of survival were similar to those provided by time-dependent models (0.04 ± 0.01 and 0.91 ± 0.02, respectively).
To evaluate spatial variation in apparent survival within the Baltic region (H2), we added the categorical variable “country” to the initial model. The more complex model indicated similar survival probabilities for older birds in all three countries and some variation in survival probabilities among countries for young birds (Table 1). However, this model (Φa2g pt AICc = 1,811.2) was less well supported by the data compared with the initial model (Φa2 pt AICc = 1,809.8), suggesting a lack of significant variation in survival within the Baltic population.
Table 1. Probabilities (± SE, 95% confidence intervals in brackets) of re-sighting and survival of Black Storks Ciconia nigra fledged in the three Baltic States during the first year and at a later age
To study sex-specific differences in survival (H3), the variable “sex” was included. The model with sex (Φa2s pt) had higher support (AICc = 680.1) than the model without this parameter (Φa2 pt AICc = 688.6; note that the lower sample size resulted in lower AICc values than those presented above). There was a three-fold difference in apparent survival in 1y females (0.013 ± 0.006; 0.006–0.030) compared with that for 1y males (0.045 ± 0.010; 0.029–0.070); the CIs were nearly exclusive (Figure 2). In older birds, we did not detect a significant difference in apparent survival between males (0.821 ± 0.035; 0.742–0.879) and females (0.784 ± 0.073; 0.609–0.894). Estimated real survival of GPS-tracked adult males was 0.80 ± 0.05 (0.69–0.88) and that among females was 0.70 ± 0.10 (0.47–0.85); however, the CIs overlapped extensively, indicating a non-significant difference between the sexes. Unfortunately, our data were too limited to obtain sex-specific estimates for young birds. Natal dispersal distances were significantly longer in females (189.1 ± 193.9 km, mean ± SD, n = 17) than that in males (72.0 ± 53.7 km, n = 27; W = 372, P <0.001). However, there was no difference between dispersal directions (females: 238.2 ± 79.0 degrees, males: 184.6 ± 107.6 degrees; W = 3.56, P <0.168) as both sexes generally dispersed southwards, on average.
Figure 2. Apparent survival probabilities for male and female colour-ringed Black Storks Ciconia nigra during their first year (A) and later (B). Note the different scales on the y-axes.
To evaluate whether survival declined along with the declining abundance (H4), we compared model Φt pt with the constant model Φ.p. and the model with the constant survival and time-dependent re-sighting probability Φ.pt. The model with the time-dependent probability (Φt pt) received higher support (AICc = 1,182.8) than those of the other two models (AICc = 1,414.0 and AICc = 1,268.1). Annual estimates fluctuated during the first half of the study period, when the re-sighting probability was low; however, we noticed a steady decrease in apparent survival after the early 2000s (Figure 3). The trend between 2001 and 2020 was negative, although it did not exceed the significance threshold (β = –0.016 ± 0.009, P = 0.1).
Figure 3. Temporal dynamics of the re-sighting probability (A) and apparent survival (B) of Black Stork in the Baltic region in 1991–2020.
Discussion
Re-sightings of colour-ringed birds provided data on apparent survival in the Black Stork population east of the Baltic Sea and GPS-tracking provided comparative data for estimating real survival. Contrary to our first two hypotheses (H1 and H2), we did not detect significant differences in the apparent survival of older birds along the north–south axis. However, apparent survival of 1y birds was much lower than that of older birds and remarkably lower than an earlier estimate in Hungary (Tamás Reference Tamás2011). Consistent with our third hypothesis (H3), female survival was slightly lower than male survival, with no significant differences between older males and females; however, apparent survival of 1y females was three times lower than that of 1y males. Consistent with our fourth hypothesis (H4), we detected a negative trend over time in apparent survival; however, the differences were only close to the significance threshold.
We did not find spatial differences in apparent survival along the north–south axis of c.700 km as would be expected according to the centre–periphery hypothesis (Brown Reference Brown1984). Moreover, apparent survival estimates of storks older than one year in the three Baltic countries (0.86–0.88) were similar to those obtained for a Hungarian population (0.83) located more than 600 km south in Central Europe (Tamás Reference Tamás2011). These findings suggest that the apparent survival of older birds is not related to the position of the population in the distribution range, despite the fact that previously studied demographic variables (long-term population trends and breeding success) show continuous declines towards the northern edge of the range (Treinys et al. Reference Treinys, Lõhmus, Stončius, Skuja, Drobelis and Šablevičius2008; Strazds Reference Strazds2011; Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b; Treinys Reference Treinys and Rašomavičius2021). Yet, our findings for Black Stork are not unique because several studies similarly have reported the lack of spatial trend in demographic rates (Carbonell et al. Reference Carbonell, Pérez-Tris and Tellería2003; Pironon et al. Reference Pironon, Papug, Villellas, Angert, García and Tompson2016; Treinys et al. Reference Treinys, Väli and Berganis2017; Martín et al. Reference Martín, Onrubia and Ferrer2021). Hence, the apparent survival of adult Black Storks in the Baltic region is probably not influenced by variation in: (1) the length of the migration route, which differs up to c.10% within the region; (2) local environmental conditions, including climatic and habitat factors; (3) biotic interactions, such as competition or predation (Niedzielkski and Bowman Reference Niedzielski and Bowman2014); (4) anthropogenic disturbance. Instead, equal apparent survival rates in breeding grounds across geographical locations support the relative importance of environmental conditions in the wintering grounds on survival, which has been earlier observed both in Black Stork and White Stork Ciconia ciconia (Schaub et al. Reference Schaub, Kania and Köppen2005; Cano et al. Reference Cano, Franco, Doval, Tores, Carbonell and Tellería2013; Martín et al. Reference Martín, Onrubia and Ferrer2021).
Apparent survival estimates for older colour-ringed Black Storks (0.86–0.88) were somewhat higher than real survival (0.77), estimated by GPS telemetry. However, the difference between results based on ring re-sightings and GPS tracking was not significant as the CIs overlapped. Considering the similar estimates of survival, we conclude that adult survival rate cannot explain the decline in the Baltic population.
However, re-sightings of colour-ringed birds suggested that apparent survival is significantly lower (0.03–0.05) in Baltic 1y birds. These figures are several times lower than those estimated in Hungary using a similar method (0.16; Tamás Reference Tamás2011). The real survival estimate was somewhat higher in our study (0.11). This difference highlights the necessity to complement the analysis of ring recoveries by less biased techniques such as GPS tracking. Furthermore, for long-lived species with a long maturation period, nest monitoring could be complemented with other types of monitoring (e.g. monitoring known feeding sites) to increase the re-sighting probability of young individuals and thus improve the reliability of survival estimates. However, both obtained figures were exceptionally low in comparison with those for other long-lived birds (Newton et al. Reference Newton, McGrady and Oli2016). As an example, only four out of 27 GPS-tracked juveniles reached the age of one year, and only one of them reached breeding age (and started breeding; M. Strazds, unpublished data). Hence, our results indicate the exceptionally high mortality rates of young Black Storks.
There are several potential explanations for the particularly low survival of Baltic Black Storks during their first year. Deaths seem to start during the late nestling period but not in the earlier phases because the clutch size in the Baltic States is similar to those in other regions (Janssen et al. Reference Janssen, Hormann and Rohde2004), and a peak in mortality is not detected during the first weeks after hatching in Black Stork (Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b; M. Strazds, unpublished data), whereas it is found in White Stork (Kosicki Reference Kosicki2012). Instead, survival likely decreases later. For instance, Väli et al. (Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b) confirmed the death of eight of 14 Black Stork nestlings during the late nestling period and first days of migration and a high mortality shortly after fledging has been recorded in Latvia (M. Strazds et al., unpublished data).
There are several mutually non-exclusive explanations for the low survival of 1y birds. First, the condition of large nestlings and fledglings may be poor due to a food shortage. We suspect that this may be the result of forest drainage, which affects biodiversity in watercourses (Rosenvald et al. Reference Rosenvald, Järvekülg and Lõhmus2014) and reduces the abundance of Black Stork prey. This may have acted in concert with climate change, particularly increased temperatures and decreased precipitation in July and August, the hottest summer months in northern Europe, when young storks fledge. Secondly, Baltic Black Storks are affected by recent DDT contamination, which has likely increased mortality among nestlings and fledglings (Strazds et al. Reference Strazds, Bauer, Väli, Kukāre and Bartkevičs2015). Thirdly, Black Stork nestlings and fledglings may suffer from predation. Although Konovalov et al. (Reference Konovalov, Nellis, Nellis, Nurmla, Sellis and Väli2019) did not detect any significant impact of predation, recent records from webcams and camera traps suggest that pressure by avian predators on young Black Storks is increasing (authors’ pers. obs., see also Langgemach and Henne Reference Langgemach and Henne2001; Zawadski et al. Reference Zawadzki, Zawadzki, Drozdowski and Zawadzka2022).
Among 1yr birds, apparent survival was significantly lower in females than in males. This could be explained in part by a higher natal dispersal of females, a characteristic pattern in many long-lived birds (Faccio Reference Faccio2013; Murphy et al. Reference Murphy, Stahlecker, Millsap, Jacobson, Johnson and Smith2018; Morandini et al. Reference Morandini, Muriel, Newton and Ferrer2019; Väli et al. Reference Väli, Kalvāns and Tuvi2021a), and also detected in the current study. Similarly, White Stork females settle at significantly longer distances from natal sites than corresponding estimates for males (Chernetsov et al. Reference Chernetsov, Chromik, Dolata, Profus and Tryjanowski2006). Furthermore, birds migrating through regions with suitable nest sites and potential partners may settle down en route instead of continuing migration (Chernetsov et al. Reference Chernetsov, Chromik, Dolata, Profus and Tryjanowski2006; see also Itonaga et al. Reference Itonaga, Köppen, Plath and Wallschläger2010). Indeed, the studied Black Storks generally dispersed southwards. Hence, natal dispersal may explain the low rates of female apparent survival and eventually disproportionately reduce populations in the northern part of the range.
However, we believe that sex-dependent mortality also shaped the sex difference detected in the apparent survival of young Black Storks. Kenward et al. (Reference Kenward, Marcström and Karlbom1999) reported markedly lower survival rates for 1y males (smaller sex) than 1y females (larger sex) in Eurasian Goshawk Accipiter gentilis. Among various explanations for this pattern, the authors proposed that the small size of males may have provided lower body reserves than those of larger females for survival during periods of food shortage. We supposed that a food shortage is the main cause of the low survival of young Black Storks in the Baltic States (see above), in which case we could also predict that survival rates will be lower in the smaller sex. Indeed, the body mass of fledgling females is lower than that of males (2.7 kg vs 3.2 kg at 50 days of age, respectively, n = 158, M. Strazds, unpublished data), corresponding to the apparent dimorphism in adulthood (males are slightly larger than females; Cramp and Simmons Reference Cramp and Simmons1977). This size difference could be particularly important during the first migration of inexperienced Black Stork juveniles. In long-lived species, distances travelled and time spent at stop-overs are longer for juveniles than adults (Sergio et al. Reference Sergio, Tanferna, De Stephanis, Jiménez, Blas and Tavecchia2014), consistent with observations of Black Storks, in which juveniles were observed in stop-over sites in Israel significantly later in the autumn compared with observations of adults (ringing data from Lithuania; R. Treinys, unpublished). Irrespective of the reason for the relatively low apparent survival in young female Baltic Black Storks, this may skew the adult sex ratio in the population. Higher female mortality is observed in many bird species but is more pronounced in threatened species or populations (Donald Reference Donald2007). At the northern margin of the range, an increasing lack of partners is an important threat for Black Stork and probably an essential driver of the population decline (Konovalov et al. Reference Konovalov, Nellis, Nellis, Nurmla, Sellis and Väli2019; Väli et al. Reference Väli, Nellis, Kaldma, Vainu and Sellis2021b).
We noticed a declining (although not significant) trend in apparent survival in Black Stork. Unfortunately, our sample sizes were not sufficient for analysing temporal trends separately in the two age classes. However, in White Stork, declines in survival in both age classes have been detected (Martín et al. Reference Martín, Onrubia and Ferrer2021). Even a non-significant trend is alarming in declining populations, as survival often has a great impact on population dynamics, especially in long-lived species (Doherty et al. Reference Doherty, Schreiber, Nichols, Hines, Link and Schenk2004; Schorcht et al. Reference Schorcht, Bontadina and Schaub2009; Sandercock Reference Sandercock, Murray and Sandercock2020), emphasising the need for further monitoring and conservation measures.
Conclusions
The most important finding of the current study was the alarmingly low apparent survival of Baltic Black Storks during their first year of life, especially among females. This, along with the declining population numbers, strongly questions the viability of the population. The most plausible explanation is a shortage of food, along with freshwater contamination with environmental pollutants, predation, and climate change. Therefore, further research should evaluate the causal relationship between these potential factors and the survival of young Black Storks in the Baltic region. If a food shortage is a key factor in explaining the low survival of young birds, then supplementary feeding could be considered as an additional conservation measure in the short term, as this approach has been successful in restoring threatened populations (Janssen et al. Reference Janssen, Hormann and Rohde2004; Jones Reference Jones, Sutherland, Newton and Green2004). However, this measure is labour-intensive, influences only a fraction of the population, and can also have negative consequences (e.g. Carrete et al. Reference Carrete, Donázar and Margalida2006). In the medium term, the building of water reservoirs in drainage systems can improve the survival of prey during the summer when drainage ditches are dried up. In the long term, the environmental quality of managed forests should be recovered by restoring damaged watercourses (brooks and streams).
Acknowledgements
Raivo Endrekson, Rein Nellis, Renno Nellis, Ain Nurmla, Urmas Sellis and many other people helped to ring Black Storks and collect re-sighting data in Estonia. The following people have assisted in colour ringing and the installation of camera traps in Latvia: Andris Avotiņš (jun.), Madars Bergmanis, Aivis Berents, Agris Celmiņš, Pēteris Daknis, Helmuts Hofmanis, Imants Jakovļevs, Eduards Kalnāre, Sandis Kājiņš, Rolands Kāpostiņš, Jānis Ķuze, Edgars Lediņš, Andis Liepa, Torkel Lundberg (Sweden), Harijs Meiers, Valdis Muižnieks, Thierry Mulders (Belgium), Eduards Ozoliņš, Lubomir Peške (Czech Republic), Aivars Petriņš, Valters Pranks, Atis Rastiņš, Ģirts Strazdiņš, Jānis Urtāns, Krister Wahlström (Sweden), Pierre Warlomont (Belgium), and Jan Wester (Sweden). Deivis Dementavičius, Darius Musteikis, Darius Norkūnas, Daiva Norkūnienė, Saulis Skuja, Darius Stončius, and many other ornithologists and foresters contributed to the Black Stork fieldwork in Lithuania. Recovery data were collected by Estonian, Latvian, and Lithuanian Bird Ringing Centres with support from Rein Nellis and Olavi Vainu. Hans-Günther Bauer, Wolfgang Fiedler, and Urmas Sellis contributed to GPS tagging, and Urmas Sellis also helped to collate the information on mortality of GPS-tagged birds. Comments of three anonymous reviewers improved the manuscript. The study was financed by the Estonian Environmental Board, Estonian Environmental Investments Fund (grant no. 17341, Estonian University of Life Sciences (project no. P180271PKEL), Latvian Council of Science (grant no. lzp-2020/1-0005), and the University of Latvia (project “Development of the science infrastructure of the national research center for forest and water resources”).
Open access
