Large Increase in Nest Size Linked to Climate Change

XXX

Large increase in nest size linked to climate change:

An indicator of life history, senescence and condition

Anders Pape Møller*,1 . Jan Tøttrup Nielsen2

1 Laboratoire d'Ecologie, Systématique et Evolution, CNRS UMR 8079, Université Paris-Sud, Bâtiment 362, F-91405 Orsay Cedex, France;

2 Espedal 4, Tolne, DK-9870 Sindal, Denmark

Word count: 6194

Figures: 3

Tables: 3

Correspondence to APM:

Tel: (+33) 1 69 15 56 88

Fax: (+33) 1 69 15 56 96

E-mail:

Author Contributions: APM and JTN conceived and designed the study. JTN collected the data. APM analyzed the data and wrote the manuscript.
Abstract Many animals build extravagant nests that exceed the size required for successful reproduction. Large nests may signal the parenting ability of nest builders suggesting that nests may have a signaling function. In particular many raptors build very large nests for their body size. We studied nest size in the goshawk Accipiter gentilis, which is a top predator throughout most of the Nearctic. Both males and females build nests, and males provision their females and offspring with food. Nest volume in the goshawk is almost three-fold larger than predicted from their body size. Nest size in the goshawk is highly variable and may reach more than 600 kg for a bird that weighs ca. 1 kg. A fraction of 8.5% of nests fell down, but smaller nests fell down more often than large nests. There was a hump-shaped relationship between nest volume and female age with a decline in nest volume late in life as expected for senescence. Clutch size increased with nest volume. Nest volume increased during 1977-2014 in an accelerating fashion, linked to increasing spring temperature during April, when goshawks build and start reproduction. These findings are consistent with nest size being a reliable signal of parental ability, with large nest size signaling superior parenting ability, senescence, and with nest size indicating climate warming.

Keywords: Accipiter gentilis . Climate change . Extended phenotype . Goshawk . Senescence . Signal

Introduction

Many animals build nests that are structures used for holding offspring (Collias and Collias 1984; Hansell 2000, 2007). Nest building has evolved numerous times independently in taxa as diverse as arachnids, crustaceans, insects, fish, reptiles, birds and mammals (Hansell 2007). Nests are often large and elaborate structures that may reach a height of several meters as in termites and ants and weigh several tons (Hansell 2007). The excessive amount of time and energy sometimes invested in nest building begs the question as to what are the advantages of such large nests. The traditional answer is that they function in protection and rearing of offspring although minimal structures required for efficient parental care can hardly explain exaggerated nest structures in many other taxa. Indeed, nest size is twice as large in birds species with biparental nest building compared to species with uniparental building, even when their clutch size is the same (Soler et al. 1988).

Nests constructed early during spring, at high latitudes and at high altitudes where the weather is colder protect offspring against inclement weather by being larger and better insulated (Schaefer 1980; Kern and Van Riper 1984; Møller 1984; Mainwaring and Hartley 2008; Crossman et al. 2011; Mainwaring et al. 2012). Such benefits of insulation may be traded against the higher risk of predation because large nests are more conspicuous and hence suffer more from predation (Møller 1990). Alternatively, specific types of nest material may provide protection of eggs and nestlings against bacteria and parasites (Wimberger 1984; Mennerat et al. 2009; Peralta-Sánchez et al. 2010). An optimal nest size will prevent excessive fouling of nests and the associated fitness costs of nestling death by allowing parents to keep the nest clean, as demonstrated experimentally for starlings Sturnus vulgaris (Erbeling-Denk and Trillmich 1990). Improved insulation provided by larger nests may benefit parent birds by preventing heat loss from eggs and nestlings. Large nests may reduce crowding of the offspring and prevent them from falling out of the nest (Slagsvold 1982, 1989; Heenan and Seymour 2011). Thus, reproductive success is positively correlated with nest size (Alabrudzinska et al. 2003; Álvarez and Barba 2008; Møller et al. 2014a, b). Nests may act as signals revealing features of the quality of the nest builder just as secondary sexual characters act as signals (Tortosa and Redondo 1992; Soler et al. 1998a, b; Møller 2006; Broggi and Senar 2009; Sanz and García-Navas 2011; Tomás et al. 2013). This hypothesis is supported by correlational and experimental studies showing that larger nests are favored because they result in differential parental investment by the partner (Soler et al. 1998b, 1999, 2001; de Neve et al. 2004; Tomás et al. 2013). Nests are extended phenotypes that reliably reveal information about quality features of nest builders. For example, bird species with more elaborate nests have a larger cerebellum compared to species with simpler nests (Hall, Street and Healy 2013). Thus nest structures may reliably reflect features of the cognitive abilities of the nest builder.

The objective of this study was to analyze long-term data on nest size in the goshawk Accipiter gentilis. We tested a number of predictions relating to the hypothetical signalling function of large nests using the goshawk as a model system. Nest size may act as an indicator of phenotypic quality as described above. If nests are larger than predicted for containers that can safely hold eggs and nestlings, we would expect that (1) nest size is larger than predicted from the allometric relationship between nest size and body size in birds. If nest size reflects individual reproductive ability, we should expect that (2) nest size is significantly repeatable among females and more repeatable than the effect of locality on nest size. If large nest size reflects superior quality, we should expect that (3) smaller nests were more likely to fall down, while the null hypothesis would be that a large and unwieldy nest would be more likely to fall to the ground. If the nest building ability of females is reflected in larger nests, but also greater reproductive investment, we should expect (4) clutch size to increase with nest size.

Senescence reflects the deterioration in phenotype at old age due to accumulation of mutations, disposable soma or deteriorating anti-oxidant capacity (e. g. Nussey et al. 2013). There are no previous studies investigating senescence effects in nest building. Senescence should be particularly visible for the most costly phenotypic characters. Hence nests of goshawks that may weigh several hundred kilograms may constitute a prime candidate for senescence. If nest building ability increases with age and subsequently decreases as a consequence of senescence, we should expect (5) a hump-shaped relationship between nest size and age.

Finally, nest size may increase when there are benefits to be gained from insulation of nest contents (Schaefer 1980; Kern and Van Riper 1984; Møller 1984; Mainwaring and Hartley 2008; Crossman et al. 2011; Mainwaring et al. 2012). In contrast, nest size may increase with climate warming if increasing temperatures reduce the costs of nest construction. In that situation we should expect nest size to increase with increasing temperature. Therefore, (6) nest size should increase with increasing spring temperature during a period of climate warming. There are to the best of our knowledge no previous studies of this prediction.

Many species of raptors are suitable for studies of the factors affecting nest size, especially because nest sites are used repeatedly during long periods of time, the identity of individuals can be determined from feather coloration, and nest size is extraordinarily large for species of a given body size (Cramp and Simmons 1980; Kenward 2006; Newton 2010).

Materials and methods

Study species

The goshawk is a solitary, socially monogamous large raptor distributed across most of the Nearctic (Cramp and Simmons 1980; Kenward 2006). The large nest is built 10-20 m above ground in a large tree, and it is used in successive years although pairs may have two-three nests that are used in different years. The nest consists of twigs and branches lined with twigs with green needles and after leafing green twigs from larch Larix larix and deciduous trees (Cramp and Simmons 1980; Kenward 2006). It is constructed by both sexes although males contribute disproportionately, especially to new nests (Holstein 1942). Nest size reaches up to 100 cm in diameter and 100 cm in height (Cramp and Simmons 1980), which equals more than one cubic meter and weighs more than half a tonne, assuming that the specific gravity is less than 0.50 (Wikipedia: http://en.wikipedia.org/wiki/Specific_gravity on 30 January 2015). Nest building starts up to 40 days before laying, and new material especially green material is added throughout the incubation and nestling periods. Clutch size is usually 3-4 eggs that are incubated by the female while the male provisions the female and small nestlings with food (Kenward 2006).

Methods for recording nest size

We retrieved information on nest height, nest diameter and body mass for all species of birds from Cramp and Perrins (1977-1994). Nest volume was calculated from the equation of an ellipsoid. The allometric relationship between log10-transformed nest size and log10-transformed body mass was used to calculate the expected nest size for a species with the average body mass of a goshawk (1140 g; Cramp and Simmons 1980).

Field methods

Jan Tøttrup Nielsen (JTN) systematically visited more than 120 localities that once held nests of goshawks in Northern Vendsyssel (57°10’ - 57°40’N, 9°50’ - 10°50’E), Denmark during April-August 1977-2014. This was part of a long-term population study. The study area is mainly agricultural habitats with mixed forests, villages and open heath. See Nielsen and Drachmann (2003) and Nielsen and Møller (2006) for a detailed description of the study areas. Each nest was visited at least three times during the breeding season. Not all pairs produced nestlings explaining the reduction in sample size from egg laying until the nestling period. A total of 569 nests built by 317 females were followed in this study. Not all nests were measured for logistic reasons since some nests were impossible to measure without endangering the measurer or the nest.

Goshawk nest size

JTN measured nest dimensions with a tape measure to the nearest cm recording maximum nest height, width and breadth when visiting the nest when nestlings were 18-30 days old.

Goshawk reproduction

JTN checked all nest sites of goshawks. Once having identified occupied nests these were all visited when nestlings were predicted to be 18-30 days old in an attempt to eliminate nest abandonment due to disturbance. We recorded wing length of the most developed nestling at the first visit to the nest, assuming that the first egg was laid 41 days before the first egg hatched (Holstein 1942). Hatching and laying date of the first eggs were subsequently determined based on wing length.

The identity of breeding females was confirmed from individual color, shape, length and pattern on their primaries and rectrices (Opdam and Müskens 1976; Kühnapfel and Brune 1995; Nielsen and Drachmann 2003; Kenward 2006). JTN kept all primaries and rectrices throughout the study, and these were used for subsequent identification of individual adults. Because all territories were visited annually, the age of reproducing females was estimated from the first year when a female was present in a territory, as yearlings, two years old, or at least three years old depending on plumage characters (Kenward 2006). Because the last category of individuals was at least three years old, we cannot be sure that this category did not include some individuals that were four years old. Hence age was the minimum age for these few individuals. These age estimates were corroborated by comparison with age as determined from 60 adults that had been ringed as nestlings and were later captured at the nest sites using standard traps. All these adults were correctly aged (3 one-year old, 12 two-years old, 9 three-years old, 7 four-years old, 6 five-years old, 3 six-years old, 6 seven-years old, 4 eight-years old, 3 nine-years old, 3 ten-years old, 1 eleven-years old, 1 thirteen-years old, 1 fourteen-years old, and 1 sixteen-years old). A total of 34 captured adults were not ringed as nestlings in the study area and hence must have immigrated from elsewhere. Finally, we emphasize that a total of only 16 individual goshawks that changed breeding forest during their lifetime were identified based on individual color, shape, length and pattern on their primaries and rectrices or from capture (three females). Although our study area of several hundred square kilometers allowed for tracking of individual females between territories, we cannot exclude the possibility that a few females dispersed outside our study area. Females of unknown age at the start of the study were excluded from analyses of age effect, ensuring that only females of known age contributed to the dataset.

Nest fate

Nest fate was recorded during visits as either ‘retained’ when the nest was still present, ‘fallen down’ when the nest was located on the ground below the nest tree, ‘cut down’ when nest trees or entire woodlots were cut, or the nest was ‘shot down’ as evidenced by gunshot.

Climatic conditions

Monthly mean temperatures from Aalborg were downloaded from the Danish Meteorological Institute. We have previously used winter (December–February) temperatures as a measure of winter severity (Herfindal et al. 2015), which is a period of the year when food scarcity may affect female condition and energy reserves needed for breeding. March temperature represents a measure of the period during initiation of breeding. April is the main period of egg laying and incubation during which climate can play a major role for reproductive success in birds (e.g. Herfindal et al. 2015; Nord et al. 2010 and references therein), whereas most eggs have hatched and environmental conditions can be important for chick survival in May.