Introduction

The onset of breeding in birds is determined largely by the food availability and prevailing weather conditions (Perrins 1970; Drent and Daan 1980; Korpimäki 1987; Daan et al. 1990; Newton 1998; Dunn 2004; see also Tökölyi et al. 2012). In particular, the rise in the ambient temperature after winter appears to be a common key factor (von Haartman 1963; Lofts and Murton 1968; Crick et al. 1997; Sokolov and Payevsky 1998; Dunn and Winkler 1999; Svensson 2004; Both et al. 2005a). This implies that the recent rise in spring temperatures in northern latitudes, which indicates global warming, had led to the concurrent earlier onset of breeding, particularly in certain income breeders (e.g. Coppack and Both 2002; Both et al. 2004, 2005b; Dunn 2004; Svensson 2004; Schaefer et al. 2006). The timing of breeding in income breeders such as insectivorous passerines is governed by the food supply in a particular environment during the potential breeding season. The advancement of breeding is not as apparent in capital breeders, such as birds of prey, in which timing of breeding is governed largely by food supply and energy stores accumulated before the breeding season (Daan et al. 1989; Jönsson 1997; see also Houston et al. 2007; Stephens et al. 2009). Nevertheless, their food supply may also be affected by habitat (e.g. Aparicio and Bonal 2002; Chace and Walsh 2006) and variable climatic conditions (e.g. Newton 1998; Both 2010).

The early onset of breeding has been suggested as a favourable trait that evolved to ensure the best feeding conditions for the forthcoming young (e.g. Lack 1968; Perrins 1970; Drent and Daan 1980; Daan et al. 1989, 1990; Verboven and Visser 1998; Blūms et al. 2005) or, typically in various long-lived species of northern latitudes, to ensure that the time-consuming reproduction process can be conducted within the strict limits of tolerable environmental conditions (e.g. Lepage et al. 1996). The latter is particularly important for young birds to successfully reach independence (Newton 1979; Pietiäinen 1989; Grandāns et al. 2009). Early breeding may also offer the opportunity to produce more clutches and young within a season (Fredrickson and Hansen 1983; Korpimäki 1987; Dowding et al. 1999; Blomqvist et al. 2001; Kloskowski 2001; Halupka et al. 2008). Together, these factors mean that the ability to respond quickly to climatic conditions can be considered profitable (e.g. Berthold 1990; Prop and Deerenberg 1991; Sergio 2003). However, sudden shifts to earlier breeding may be detrimental if the critical periods for successful reproduction are changed to less suitable ones (Visser et al. 1998, 2004; Sanz 2003; Both et al. 2005a, 2009; Both and Visser 2005). For instance, it may be possible to advance laying dates to the extent that most or all individuals in a population breed too early (e.g. Cresswell and McCleery 2003).

The Tawny Owl Strix aluco is a widespread bird of prey of rural and urban habitats in Europe (Mikkola 1983; Cramp 1985). It commonly occupies habitats in the vicinity of human settlements including near the northern limit of its range in Finland, which was only colonized approximately 100 years ago (Saurola 1995; Solonen 2005). A suitable nesting cavity is an essential prerequisite for the Tawny Owl’s breeding, which is otherwise strictly governed by an adequate availability of food, usually arvicoline voles (Linkola and Myllymäki 1969; Solonen and Karhunen 2002; Kekkonen et al. 2008). The availability of food for breeding Tawny Owls is particularly affected by the pronounced and more or less cyclic fluctuations of vole populations (Solonen 2004, 2010; Sundell et al. 2004), and is modified by habitat (Petty 1989; Solonen and af Ursin 2008) and various climatic factors such as winter temperature and snow cover (Southern 1970; Solonen and Karhunen 2002; Solonen 2004, 2005; Lehikoinen et al. 2011). Generally speaking, the food supply for some species (such as some tits and birds of prey) seemed to be higher in urban habitats than in rural ones, particularly during winter and before the breeding season (Marzluff et al. 2001; Chace and Walsh 2006; Chamberlain et al. 2009). For Tawny Owls, urban habitats seem to be particularly favourable owing to the abundant and relatively stable availability of prey that commonly accompany human settlements (such as various small mammals and birds) as well as the prevailing benign microclimate (Solonen and af Ursin 2008). Mild winters also seem to be favourable, but they may also have negative effects in terms of their variable impact on small mammals (Solonen 2004, 2006). In the heavily fluctuating environmental conditions of the north, breeding of Tawny Owls seem to be most frequent and most successful in territories that provide the most stable year-round food supply (Solonen and af Ursin 2008). In such territories, breeding might also begin earlier than elsewhere. Due to relatively stable environmental conditions, Tawny Owls may even breed throughout the year in some areas of southern Europe (Zuberogoitia et al. 2004).

I examined the timing of breeding in rural and urban Tawny Owls of the capital (Helsinki) district and its surroundings in southern Finland between 1986 and 2011. My primary goal was to assess the role of fluctuating vole abundance and climatic conditions of the preceding winter in the timing of breeding, as well as to highlight the differences between the main habitat types of the study area (rural vs. urban). I expected that (1) Tawny Owls bred earlier in urban habitats than in rural ones, due to more stable and benign environmental conditions, (2) high abundance of voles and mild and less snowy winters advanced the onset of breeding, and (3) the effects of fluctuating vole abundance and winter weather between rural and urban habitats differ, probably due to the additional availability of less variable prey populations and to less frost and snow in urban areas. An alternative explanation could be that urban habitats locally dampen the general vole fluctuations. In particular, I expected that the availability of other prey than small voles was higher in urban areas than in rural ones, and therefore the level of fluctuating abundance of voles needed to activate breeding in owls was lower in urban than in rural habitats.

Methods

Study on Tawny Owls

The present study was conducted in Uusimaa (60°N, 25°E), near the southern coast of Finland (Solonen and Karhunen 2002) between 1986 and 2011. The study area of more than 500 km2 consisted of low-lying rural habitats of mixed fields and forests, as well as the capital district (Helsinki and its surroundings), the most urban area of the country. Tawny Owls occupied the study area in a relatively even manner, following the availability of suitable nest-sites, mainly nestboxes erected throughout the study area. The hatching success of birds was recorded for each year of the study. Only the successfully hatched clutches were included in the present data. Possible biases caused by variable hatching success should be minor because unhatched eggs were rare in successful nests (in nests from which at least one young fledged).

In order to estimate the hatching dates, I measured the wing length of nestlings or recently fledged young from 554 broods using the maximum method (e.g. Svensson 1984), from the carpal joint to the tip of the wing. I estimated the age of the oldest young using Jokinen’s (1975) wing-growth curve. The hatching dates were estimated backwards from the age of the young. According to Jokinen (1975), the growth of the wing length in Tawny Owls is rather even and independent on the variations in the food supply. During the first week after hatching, the age of nestlings can be estimated on an accuracy of 1–2 days based on the wing length. Later, the variation somewhat increases but the accuracy is still 3–4 days even during the last half (2 weeks) of the nestling period. For the timing of egg laying, I used an incubation time of 28 days (Cramp 1985; Saurola 1995). Calendar days were converted to Julian days, starting from the first of January.

Habitat, food and weather variables

In order to demonstrate the effect of two main habitat types, the 179 locations of nesting territories were classified into rural (146) and (sub)urban (33). The rural territories were situated in agricultural and silvicultural environments, while the suburban and urban territories were located in recreational forests and parks of built-up areas (Solonen 2008b; Solonen and af Ursin 2008). This rough categorisation seems suitable for the present purpose because the actual hunting areas of the owls were unknown (cf. Solonen 2008b). In order to demonstrate the effects of local habitats, I gave each nesting territory an individual identity number (ID).

In order to characterise the general level of food supply for owls before the breeding season, I used the regional small vole indices of the preceding autumn ("Appendix"). They seemed to be associated with the occurrence and nesting of Tawny Owls better than the respective spring indices that instead reflect the vole abundance during the nestling period of owls (Solonen 2010). Each index was based on a total of 384 trap-nights of trappings conducted during 2-day periods in early October at two localities near the present study area and included both forest and old field habitats, 192 trap-nights each (Kimpari Bird Projects; Solonen and Ahola 2010; Solonen 2010). Thus, they reflected only general levels of fluctuations in numbers of voles in the district, not particularly in the present rural and urban habitats. These indices seemed to characterise fluctuations in the abundance of voles over a larger area in southern Finland relatively well (Solonen 2004, 2010). The vole species concerned were the Bank Vole Myodes glareolus and the Field Vole Microtus agrestis. The Field Vole is the most common prey species of breeding Tawny Owls in the study area (Solonen and Karhunen 2002).

I also examined the effects of local mean winter temperatures, including January, February and March, as well as the depth of the snow cover in the middle of March ("Appendix") measured at the Helsinki-Vantaa airport (Finnish Meteorological Institute), roughly in the middle of the present study area. These variables seemed to reasonably characterise the strength of winter before the Tawny Owls’ breeding season. Preliminary examination of the data showed that there was a significant negative correlation between the variables (r = −0.693, P < 0.001, df = 24).

Statistics

Trends and other relationships between pairs of variables were examined by least-squares linear regression and the non-parametric Mann–Whitney U test was used for assessing whether hatching dates tend to be earlier in urban than in rural habitats, using VassarStats statistical software (http://www.vassarstats.net/). In order to obtain a model, which combined the effects of various explanatory variables, I used mixed effects models (lme) (Pinheiro and Bates 2000) from nlme in the R statistical package (R Development Core Team 2008; Venables et al. 2008). In the analyses, I used the Julian hatching date of the largest (i.e. oldest) young in the brood as the response variable. The explanatory variables (fixed effects) included the general habitat (rural vs. urban), the regional vole abundance index of the preceding autumn, and the weather variables of the current/preceding winter (mean winter temperature and snow cover in March) ("Appendix"). Separate and joint effects of variables were both considered. The territory ID, indicating impacts of the local habitat and individual birds, and year, indicating annually varying unknown factors, characterised random effects. The general structure of the model was as follows: hatching date ~ habitat + voles + weather + habitat × voles + habitat × weather + voles × weather, random = ~hatching date|territory × hatching date|year.

Results

The hatching dates of Tawny Owls varied between Julian dates 21 and 151 (median 119, n = 554). The hatching dates of the earliest broods of the season varied annually between Julian dates 21 (2009) and 135 (1987) and the respective variation in the annual median dates ranged from 100 (2009) to 143 (1996) (Fig. 1). The length of the laying period (days between the earliest and latest hatching date of season) varied between 9 and 122 days, with a median of 45 days. There was no significant trend (based on least-squares linear regression) either in the earliest or the median hatching dates of the population during the study period (Fig. 1; the earliest clutches: t 24 = 0.296, P = 0.770; medians: t 24 = 0.135, P = 0.894). This also applied to the annual explanatory variables (vole indices, weather variables) studied (t 24 < |2|, P > 0.05).

Fig. 1
figure 1

Fluctuations of the earliest (lowest graph), median and latest hatching dates of the season in a Tawny Owl Strix aluco population in southern Finland between 1986 and 2011

Full size image

The range in hatching dates was 28–151 (median 120, n = 453) in rural habitats and 21–149 (median 109, n = 101) in urban ones. The median hatching dates of urban territories were significantly earlier than those of rural ones (median Julian dates 109 and 121, respectively; Mann–Whitney U test, U 146.33 = 1239.5, z = 4.35, P < 0.001) (Fig. 2). The earliest hatching dates showed a similar pattern (median Julian dates 96 and 112, in urban and rural territories, respectively; Mann–Whitney U test, U 146.33 = 978.5, z = 5.32, P < 0.001). There were no trends in the earliest or median hatching dates either in rural or in urban environments (t 24 < |2|, P > 0.05). The exceptionally early clutches were laid within two periods (1988–1995 and 2008–2009) (Fig. 1) in two urban territories. The two earliest clutches found in the present data were laid in the same urban territory in the 2008 and 2009 breeding seasons. The first eggs of the clutches were laid on 30 December, 2007 and 24 December, 2008, respectively.

Fig. 2
figure 2

Distribution of median Julian hatching dates in rural (open bars, n = 146) and urban (closed bars, n = 33) territories of Tawny Owls in a population in southern Finland between 1986 and 2011

Full size image

In association with the effect of habitat, the breeding of Tawny Owls began earlier the higher the (autumn) vole levels and the milder the preceding winter (Table 1). There also emerged a significant joint effect between the vole abundance and snow cover: owls bred earlier the more there were voles and snow simultaneously. The main effects of the increasing winter temperature and depth of the snow cover on the timing of breeding in Tawny Owls were, however, delaying.

Table 1 The mixed-effects model characterising the roles of the fixed effects, the general habitat (habitat = increasing urbanization), vole abundance (voles), mean winter temperature (temperature) and depth of the snow cover (snow) in determining (Julian) hatching dates for the clutches of rural and urban Tawny Owls Strix aluco in southern Finland between 1986 and 2011
Full size table

Discussion

Effects of habitat

Tawny Owls bred earlier in urban territories than in rural ones. High general abundance of voles and mild winters advanced the breeding of owls particularly in urban habitats. This suggests that the treshold to start egg laying was lower in urban territories, i.e. urban birds were, on average, better prepared for the onset of breeding, and a smaller improvement in environmental conditions might trigger their breeding.

Early breeding in urban habitats seems to be a widespread phenomenon (see Chamberlain et al. 2009). Exceptionally early clutches of Tawny Owls have been laid in urban habitats, both in the present northern study area (Solonen 2000; this study) and elsewhere at more southern latitudes (e.g. Petriņš 1986; Godovanets et al. 1992; Grandāns et al. 2009). Such clutches have occurred particularly in good vole years. Otherwise, the factors determining the early onset of breeding in urban environments are somewhat unclear. Possible explanations include the better general availability of food, since rodents and small passerines accumulate in urban and suburban areas in winter due to feeding and other human-provided resources (e.g. Petriņš 1986; Grandāns et al. 2009). In addition, a higher mean ambient temperature (e.g. Marzluff et al. 2001) and anthropogenic light sources in towns and cities might physiologically activate birds earlier in spring than in more natural conditions (Partecke et al. 2004; Grandāns et al. 2009).

Effects of regional vole abundance and winter weather conditions

Besides being significantly earlier in urban than in rural habitats, in association with the effect of habitat the breeding of Tawny Owls began earlier the higher the vole levels and the milder the preceding winter. These findings were in accordance with expectations. However, the main effects of both increasing mean winter temperatures and depth of the snow cover seemed to be the delayed breeding of owls. Instead, owls bred earlier the more there were voles and snow simultaneously. This suggests that deep snow cover was favourable for over-wintering voles (Solonen 2004, 2006).

Winter temperature and snow cover characterise different kinds of constraints set by winter on the breeding of Tawny Owls, namely the effects on the energy requirements of the birds themselves and the impact on the actual availability of their prey, respectively. The direct effects of the present weather variables indicating mildness of winter cover somewhat different periods of time. They may also have various indirect effects, particularly on the availability of owls’ food (Solonen 2004, 2010).

The significant contribution of benign winter weather in advancing the onset of breeding was probably partly associated with reduced energy requirements and better hunting success on small voles in less snowy winters (Solonen 2004, 2005; Grandāns et al. 2009). In particular, under snow hiding Water Voles Arvicola amphibius are earlier available after mild winters. In southern Finland, the Water Vole is, in general, the most important prey species of Tawny Owls during breeding (Solonen and Karhunen 2002; Kekkonen et al. 2008). After a mild winter, this may also be the case before breeding, which contributes to the timing of egg laying of owls. It has been suggested that when the importance of small voles is not evident, the abundance of larger Water Voles might play the most important role in the timing of breeding in Tawny Owls (Kekkonen et al. 2008). Also, the advancing effect of the interaction between the abundance of small voles and the depth of the snow cover seems reasonable. When there are voles in abundance in autumn and the snow cover is deep in winter, more voles will survive until spring and to the laying period of owls.

Has the onset of owls’ breeding advanced between years?

The Tawny Owl breeding season usually starts in about mid-March across Europe (Mikkola 1983; Petty et al. 1994; Saurola 1995; Grandāns et al. 2009). Recently, however, increasing numbers of early (winter) breeding attempts have been recorded (e.g. Petriņš 1986; Godovanets et al. 1992; Solonen 2000; this study; Owen 2003; Grandāns et al. 2009). These observations might suggest that the onset of breeding in Tawny Owls has advanced not only within the season but also within a longer term between breeding seasons. However, the present data showed no trends in the laying dates of owls nor in the explanatory variables during the study period of 26 years.

Heavily fluctuating food supply and weather conditions seem to cause occasional short-term shifts to advanced breeding among individuals of a population within a season, or in some individuals between consecutive breeding seasons. Long-term advancement of breeding may be possible if there are favourable trends in certain factors governing the timing of breeding. Recent long-term advancements in timing of breeding in birds seemed to be largely human-induced and have appeared either as spatial gradients or temporal trends. The former are largely due to urbanization (e.g. Marzluff et al. 2001; Partecke et al. 2004; Chamberlain et al. 2009), and the latter particularly due to climate warming (e.g. Møller et al. 2010). Global climate warming has also been the most commonly advocated factor behind advanced breeding in birds (Both et al. 2004; Crick 2004; Dunn 2004). However, even during relatively long periods, such as that of the present study, various climatic factors indicating temperature have not shown any trend to suggest global warming (see also Svensson 2004; Solonen 2008a). In fact, the winter index of the North Atlantic oscillation (e.g. Hurrell et al. 2001), one of the most commonly used indicators of global warming, has actually declined during the 26 years of the present study (http://www.cru.uea.ac.uk/cru/data/nao.htm; accessed 6 February 2012). In the present case, the high fluctuations of vole populations may also have blurred any potential trend in the hatching date. Elsewhere, and for longer time scales, however, increasing global or local spring temperatures (suggesting climate warming) as well as advanced breeding in birds have been documented in several cases (e.g. Walther et al. 2002; Crick 2004; Watkinson et al. 2004; Both et al. 2009; Møller et al. 2010; Schroeder et al. 2012). Therefore, the significant relationships between the hatching date and weather factors related to ambient temperature show that the timing of breeding in Tawny Owls may become earlier with warming climate.

Advanced breeding does not necessarily concern total populations, but rather single high-quality individuals, pairs or territories. This is suggested by single and repeated records from various nesting territories. Most early breeders are probably experienced, elderly birds (Blūms et al. 2002; Verhulst and Nilsson 2008; cf., however, Grandāns et al. 2009). In the present study population, an especially long-lived female frequently laid her clutches in January and advanced her breeding significantly during her reproductive career (Solonen 2000). The timing of egg laying fluctuated in accordance with local winter temperatures.

Consequences of advanced breeding

The present results suggest that favourable combinations of environmental conditions may occasionally or repeatedly advance the onset of breeding in Tawny Owls. Advanced breeding may, however, cause chicks to hatch or fledge in hostile weather conditions of early spring, when the general food availability is traditionally at its worst (Solonen 2000). So, this kind of ‘mistiming’ may reduce the benefit gained from the early breeding if the food supply does not improve with the advancement of breeding (cf. Visser et al. 2004; Schroeder et al. 2012). In most of the species studied to date, the advancement in the beginning of breeding coincided with the advancement in the end of breeding; this resulted in a shift in the entire breeding season, with its length remaining unchanged or slightly shortened (e.g. Both and Visser 2001; Sanz 2003; Schaefer et al. 2006). Only a few authors have reported lengthening of the breeding season, with similar consequences as in Tawny Owls in some years of the present study (Halupka et al. 2008; Najmanová and Adamík 2009).

After leaving the nest-hole, the mortality of young Tawny Owls may be considerable due to both mammalian predators (such as the Pine Marten Martes martes and the Red Fox Vulpes vulpes) and avian predators (such as the Northern Goshawk Accipiter gentilis and the Eagle Owl Bubo bubo) (Sunde 2005; Solonen 2011). The mortality risk may increase with the fledging date from 14 % in April to more than 58 % in June (Sunde 2005). Therefore, the increasing predation of late broods may be an important selective agent for early breeding in the Tawny Owl. However, poor food supply, adverse weather conditions and deficient cover against predators provided by vegetation very early in the spring may have a contrary effect. In timing of breeding, laying females should compromise between optimal food availability for and minimal predation pressure against their young. From a parental perspective, the optimal laying date is a trade-off between self-maintenance costs and fitness benefits for offspring (Daan and Tinbergen 1997). If females lay when the food availability for themselves is low, their reproductive investment and survival will be affected (Drent 2006).

The occasional or regular lengthening of breeding seasons might increase the occurrence of replacement nestings and might even lead to genuine second clutches in Tawny Owls. The breeding schedule of northern owls is, in general, so tight that such replacement nestings or second clutches are not possible (e.g. Mikkola 1983; Forsman et al. 1995). Southern (1970) noted that renesting in Tawny Owls occurred only in years when prey levels were high. However, renesting appears to be a usual response of diurnal raptors and owls to clutch loss during early incubation (Newton 1979; Mikkola 1983; Forsman et al. 1995). Otherwise, replacement and real second clutches in birds of prey are only known to be common in Barn Owls Tyto alba (Taylor 1994; Martínez and López 1999; Roulin 2002). In the Tawny Owl, only one probable case of a second brood after successful breeding has been reported, from Spain (Zuberogoitia et al. 2004). In the two present cases of exceptionally early breeding, in both years (2008, 2009) a new brood appeared in the territory in July (J. Laine, personal communication). Based on the developmental stage of the young, they obviously originated from nests in some of the numerous natural holes of the nesting territory, although the nests were not found. In 2008, at least two cases of renesting after the loss of eggs or nestlings were recorded in two territories in the neighbourhood (E. Haapanen, personal communication).

Concluding remarks

One may ask why the recent early clutches are so prominent in the Tawny Owl even in the present study area where the species is a relatively new inhabitant from more southern areas. The success of colonisation and breeding of the Tawny Owl in Finland is clearly largely due to the association of the species with environments that are strongly influenced by mankind. Of the northern owls, only the Long-eared Owl Asio otus resembles to some extent the Tawny Owl in this respect (e.g. Nilsson 1984). Recently, other species of northern birds of prey, particularly the Northern Goshawk and the Eagle Owl, have shown a similar attitude to man-made environments in the present study area (Solonen 2007, 2008b; Solonen et al. 2010). However, the relationship of the Tawny Owl with rural and urban habitats is much stronger than that of the other species of northern birds of prey. Rural and urban habitats, particularly the latter, provide various advantages extensively and effectively used by Tawny Owls (e.g. Solonen and af Ursin 2008). Increasingly favourable wintering conditions due to climate change and urbanization provide advantages that can make advanced breeding and repeat or even genuine second clutches possible for an increasing proportion of the population.