Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

JOURNAL OF

1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
AVIAN BIOLOGY
Article
Heavy and persistent rainfall leads to brood reduction and nest
failure in a passerine bird

Eva Maria Schöll and Sabine Marlene Hille

E. M. Schöll (https://orcid.org/0000-0003-3096-5885) ✉ (eva.schoell@boku.ac.at) and S. M. Hille (https://orcid.org/0000-0002-3680-7549), Inst. of


Wildlife Biology and Game Management, Univ. of Natural Resources and Life Sciences, Vienna, Vienna, Austria.

Journal of Avian Biology The process of reproduction is a sensitive period for bird offspring, since they are
00: 1–11, 2020 exposed to environmental conditions for several weeks. Within the long reproductive
2020: e02418
doi: 10.1111/jav.02418 period, adult birds do not only have to maintain their own body condition, they also
have to ensure nutritional needs of their nestlings. In the worst case, if weather condi-
Subject Editor: Christiaan Both tions are very harsh, breeding is not successful and fails. This suggests that weather
Editor-in-Chief: Thomas Alerstam conditions might be an important driver of breeding success. Here, we studied the
Accepted 23 April 2020 effect of weather conditions (temperature and precipitation) on the survival of 350
nests of great tits Parus major during nestling period in six years (2010–2015). We
used traditional generalized linear mixed model (binomial response variable) and a
dedicated survival-analysis program (MARK) to analyze the data. The two approaches
allowed us to highlight and compare the effect of weather conditions on a fine (nest-
ling survival) and coarse (nest fate) scale. Both methods showed that precipitation
explained most of the variation in individual nestling and overall nest survival during
the 15-day nestling period. Heavy and persistent rainfall did not only lead to brood
reduction, it ultimately also led to losses of the entire brood. Likely causes were the
negative effects of precipitation on thermoregulation, food availability and predation
risk. However, while reduced food availability most likely might have led to brood
reduction through selective individual nestling death, predation might have resulted
in total nest failures.

Keywords: MARK, nest survival, Parus major, precipitation, temperature

Introduction
Life history of an animal is explained by traits that on one hand affect reproduction,
like clutch size, number of broods per season, age at first reproduction and lifespan,
and on the other hand affect survival. Both reproduction and survival highly affect
population growth, but survival seems to be the strongest to model demographic pro-
cesses (Schmidt et al. 2005, Arlt et al. 2008, Rose et al. 2019), as animals can only
contribute to the local population if they survive until they start to reproduce (Stearns
1992). In birds, a physiologically demanding and critical timeframe of survival until
first reproduction is the period between hatching and fledging (Ricklefs 1969).
––––––––––––––––––––––––––––––––––––––––
© 2020 The Authors. Journal of Avian Biology published by John Wiley & Sons Ltd on behalf of
www.avianbiology.org Nordic Society Oikos
This is an open access article under the terms of the Creative Commons
Attribution-NonCommercial-NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited, the 1
use is non-commercial and no modifications or adaptations are made
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
At this stage of development, several variables can severely (e.g. weather conditions): a dedicated survival-analysis pro-
affect nestlings, but abiotic conditions (e.g. local weather con- gram (MARK, White and Burnham 1999) and traditional
ditions) and their temporal variability are assumed to have a generalized linear mixed model. Using a large sample size
major effect on nestling survival and therefore on reproduc- of 350 nests, we analyzed nest fate across the 15-day nest-
tive success (O’Connor and Morgan 1982, Morrison and ling period and divided the nestling period in three 5-day
Bolger 2002, Slobodník et al. 2013). intervals to study nestling survival. While nest survival analy-
As climate change scenarios are forecasting changes in ses only account for the effect of local weather conditions
mean temperature and precipitation (IPCC 2013), studies on complete nest fates, the effect of temperature and rain-
assessing detailed information about the nature of climate and fall could also be less dramatic, leading to brood reduction
its impact on animal populations are of high interest (Knape through selective individual nestling death rather than to the
and de Valpine 2011, Matter and Roland 2017). However, loss of the entire brood. Therefore, we assume that the results
while reference periods to calculate climatological averages of both survival analyses will be complementary rather than
cover at least 30 years (World Meteorological Organization contradictory.
2017), yearly changing weather conditions are assumed to
affect demographic parameters of local populations more
directly (Skagen and Adams 2012). Near- and long-term pro- Material and methods
jections forecast that daily ambient temperature will become
more extreme and heavy precipitation events will become Study area
more frequent in temperate regions (Stocker et al. 2013), but
little knowledge exists to which degree these extreme condi- Field work on great tits was carried out in a mixed deciduous
tions affect bird species (Martin et al. 2017). forest located in lower Austria (48°05′03″N, 15°55′24″E).
While adult birds and offspring of precocial birds are usu- The vegetation is dominated by beech Fagus sylvatica and
ally well-suited and prepared for unfavorable temperature acer Acer sp., along with some conifers (spruce Picea abies, fir
and rainfall conditions, nestlings of altricial birds are most Abies alba, larch Larix decidua). Nest boxes for great tits were
vulnerable to weather conditions, because newly hatched provided within the study area (19.5 ha) along an altitudinal
young lack highly insulating down feathers and relatively gradient ranging from 491 to 887 m above sea level. Nest
waterproof plumage. Further, nestlings cannot regulate their boxes were monitored between March and July in the years
own body temperature in the first few days after hatching 2010–2015. The number of boxes increased throughout the
(Kluijver 1951). Open nests are more affected by adverse six study years from 211 (in 2010) to 286 (in 2015). We
weather conditions than cavity nests (Mainwaring et al. therefore operated with the occupancy rate of nest boxes by
2014). Nevertheless, temperature and rainfall indirectly great tits that varied between 28.0 and 48.4%.
affect survival of cavity-nesting birds (Bordjan and Tome
2014). Parental nest visitation rates decrease with increas- Data collection – breeding data
ing amount and persistence of rainfall (Öberg et al. 2015);
thus, rainfall may negatively affect nestling growth (Keller Within our long-term study, nest boxes were frequently mon-
and van Noordwijk 1994). Foraging success might be driven itored during nest building and egg-laying period every two
by weather conditions (Newton 1998), because prey activity to three days. Once a clutch was completed and eggs in the
is reduced during cold and rainy periods (Avery and Krebs nest were warm, nests were not checked for the following 13
1984). In addition, caterpillar availability is reduced dur- days to avoid disturbing the incubating female (mean incu-
ing periods with harsh weather (Schöll et al. 2016). Thus, bation period length in our study site: 12 days). Afterwards,
regional weather conditions can have a major effect on local nests were daily revisited until the last nestling had hatched.
population dynamics, if survival of individual nestlings is Once all nestlings had hatched, the brood was assumed to
reduced or the entire brood is lost (Öberg et al. 2015). be completed (day 1). In some rare cases, when nests were
Estimates of nest survival provide a sensitive metric to checked 13 days after the assumed incubation start, the last
study the effects of differences among environmental condi- nestling had already hatched. In these cases, age of nestlings
tions (Dinsmore et al. 2002, Rotella et al. 2007), whereas could be readily and precisely determined (based on a com-
estimates of individual nestlings in a nest are a useful measure- prehensive study made in that area in the year 2009, master
ment of annual reproductive success (Armstrong et al. 2002). thesis project of J. Fingerlos).
Therefore, here, we analyzed firstly, daily nest survival rates of We monitored a total of 408 great tit nests in the years
great tit Parus major within the 15-day nestling period and 2010–2015. Since great tits are multiple brooded species and
secondly, variation in individual nestling survival through- nesting success differs in first and second broods (Kluijver
out three consecutive 5-day intervals, using data on breeding 1951), we included only first broods and nests that started
performance gathered within the 6-year study period. The within 30 days after the first egg was laid in each year in the
great tit is an excellent species for studying nest fate, since population in our analyses (Both et al. 2004, Both and Visser
it is breeding in nest boxes, which are easily monitored. We 2005). Thus, we had to exclude 51 nests because the first egg
used two methods to analyze variation in survival data by in these nests was laid 30 days after the first egg had been laid
incorporating variables and factors of ecological relevance in the specific year. In addition, we had to exclude 7 nests,

2
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
because we did not know exact date of clutch initiation. The in natural systems, where nests are usually not checked on
fate of the nests was monitored on a regular basis every five a regular basis, this kind of analyses is highly advantageous.
days (day 5, 10 and 15) after the last nestling had hatched, While this advanced method has greatly improved the esti-
but some nests failed before day 15. Therefore, sample sizes mation of nest survival itself, the approach assumes that
for the two model approaches differ. daily survival probability is constant within nesting stages
and among all nests, which is often biologically restrictive
Weather data or unrealistic (reviewed by of Jehle et al. 2004). Moreover,
the method does not allow incorporating covariates that may
Daily weather conditions were obtained from a weather explain variation in nest survival. Dinsmore et al. (2002)
station, which was located within the study site (Office of therefore developed an extension of the Mayfield method
the Federal Government of Lower Austria, Hydrology and using the program MARK (White and Burnham 1999). The
Geoinformation, official webpage: <www.noe.gv.at>). Sum method includes all the advancements of the Mayfield esti-
of precipitation (mm) was measured every 15 min in the years mation, but it also accounts for daily variation in survival
2011–2015, whereas in the year 2010, only data of daily sum rates as a function of daily varying environmental conditions.
of precipitation were available. We extracted four precipita- Therefore, daily survival of great tit nests during the 15-day
tion metrics to comprehensively study the effect of rainfall. nesting period was modelled using the programs MARK
Daily sum of precipitation (mm) was measured (Bordjan and and R statistics (< www.r-project.org >). In addition, the R
Tome 2014) and number of days with rainfall (> 0 mm), package ‘RMark’ (Laake 2018) was used, which provides an
strong rainfall (> 10 mm, Skagen and Adams 2012) and interface to use the software MARK. The terms ‘nest survival’
heavy rainfall (> 20 mm, Öberg et al. 2015) within the nest- and ‘nest success’ are used in various ways, thus, we chose
ling period were counted. To assess if the amount of precipi- the following definitions for the purpose of this paper. ‘Nest
tation during the six study years was usual or extraordinary survival’ is the probability that a nest will be successful within
for the study site, we calculated the 30-year average of precip- the period between hatching and the following consecutive
itation measured at three meteorological stations surround- 15 days (nestling period). A nest was considered successful,
ing the study area (distance 24–32 km) and compared these when a) at least one fully feathered young was observed 15
data (Auer et al. 2012) to the data measured within the study days after hatching in the nest (time of fledging), or b) the
site in the years 2010–2015. nest was undisturbed at the time of fledging, but nestlings
Ambient temperature (°C) was recorded every 15 min at had already fledged, or c) we heard fledglings nearby the nest
the weather station. To verify, if temperature data collected box 15 days after hatching. Nests were considered to have
at the meteorological station represent temperature within failed, when a) we detected clear evidence of predation e.g.
the study site, ambient temperature was measured at 40 destroyed nest material, injuries in dead nestlings or adults
nest boxes along the altitudinal gradient in the year 2015. indicating attacks by other animals, or b) nest were deserted
Between 24-Apr and 29-May-2015, data on ambient tem- due to an unknown cause and chicks had starved. ‘Daily nest
perature was collected for periods between 10 and 16 consec- survival (DNS) rate’ is the probability that a nest will survive
utive days (using iButtons: every 3 min, resolution: 0.5°C). a single day within this period.
Temperature buttons were fixed underneath nest boxes to To model nest survival during the 15-day nestling period
avoid direct solar irradiation at any daytime. Daily minimum four parameters were required: a) the first day the nest was
temperatures were extracted for a period of 24 h. These daily checked during the nestling period; b) the last day the nest
minimum temperatures measured along the altitudinal gradi- content was checked being alive; c) the last day the nest was
ent were then compared with daily minimum temperatures checked during the nestling period and d) the fate of the nest
measured at the meteorological station for the same periods (binary outcome: successful/failed) (Dinsmore et al. 2002).
(Wilcoxon rank sum test with Bonferroni correction). We did These parameters allow the estimation of nest survival prob-
not find statistical differences between the temperature mea- abilities, even if nests were not monitored from the beginning
surements. Therefore, daily minimum temperature measured of the nestling period or were controlled at various irregular
at the weather station seemed to represent these temperature intervals. The program MARK uses this information to gen-
measurements within the study site adequately. erate an appropriate encounter history (live–dead format) for
each nest covering the 15-day nestling period. Since we do
Modeling of nest survival using RMark not know the exact day a nest failed, MARK models the fail-
ure of the nest during the interval from the last day the nest
During the last decades, nest survival has been usually esti- content was checked being alive until the last day the nest was
mated based on the Mayfield method, which incorporates checked during the nestling period (Dinsmore and Dinsmore
exposure days to account for lacking data if parts of the stud- 2007, Rotella 2018). We followed the approach described by
ied nests were not monitored from the beginning of nest Dinsmore and Dinsmore (2007) and standardized the begin-
initiation (Mayfield 1961, 1975). In addition, this method ning of the nestling period among years (22-Apr). We con-
accounts for irregular visits to nests, by assuming that nest fail- verted all calendar dates to numerical dates and numbered
ures occurred midway between the last two visits. Especially all nest-check dates sequentially thereafter. We defined the

3
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
maximal period great tit nests were active, which ranged from individual nestlings, we used generalized linear mixed mod-
22 April until 19 June (59 days), regardless of year. Since els (GLMMs) with a binomial distribution and a logit link
nest age can affect daily survival rates (Grant et al. 2005), we (package ‘lme4’, Bates et al. 2014). We created three sepa-
aimed to most precisely age the nests. When the last nestling rate models, each lasting for five days during the nestling
of a clutch had hatched, we aged this nest as nest age = 1. In period: day 1–5, day 6–10, day 11–15, where day 1 again
most cases, the nest was first controlled within the nestling indicated the day, when a clutch was completely hatched.
period, when only some of the nestlings in a nest had already We used the cbind() binomial approach for the response
hatched. Thus, we checked these nests on a daily routine and variable, where we included the number of nestlings at the
when the last nestling of a nest had finally hatched; we aged start and end of each 5-day period (day 1/day 5, day 5/day
this nest by back-calculating when the nest was checked the 10, day 10/day 15). Note, that the sample size used for ana-
first time during the nestling period using negative numbers lysing nestling survival using GLMMs is lower than for the
(nest age < 1). Only in some rare cases (n = 29), nests were nest survival analyses using RMark, because we could only
visited the first time within the nestling period when the last use nests, which were exactly visited on day 1, 5, 10 and
nestling just had hatched some days before (nest age > 1). 15. In addition, since some nests already failed within the
To avoid any effect of these delayed nest checks on our nest first days after complete hatching, the sample size decreased
survival model, completion days for these 29 nests were put throughout the nestling period.
in the analyses rather than actual check dates (Rotella 2018). We used a stepwise approach for model selection for each
We included 58 individual covariates for each year, 5-day period. First, we identified, which precipitation param-
accounting for daily weather conditions (minimum daily eter explained most variation in nestling survival (step 1;
temperature 24 h, daily sum of precipitation 24 h) and nest Öberg et al. 2015). Each model included one of the precipi-
age. These time-varying covariates corresponded to a unique tation metrics, calculated for the respectively 5-day interval
daily value for each nest and were included by building a within the nestling period (sum of precipitation (Bordjan and
design matrix to model daily nest survival as a function of Tome 2014), number of days with precipitation, number of
weather conditions and nest age. Finally, DNS estimates of days with strong precipitation > 10 mm (Skagen and Adams
the best model were calculated for each study year separately. 2012), number of days with heavy precipitation > 20 mm
To compare apparent fledgling success of the six study years, (Öberg et al. 2015)), and a nested random effect (study year,
we raised the estimates of DNS to the power of 15 (DNS15), nestbox ID). The four preliminary models were ranked by
since nests were monitored during the nestling period from Akaike’s information criterion (AICc) to determine the model
the hatching of the whole brood until day 15 after hatch. In with the best fit for the next step in the analyses.
total, a series of 13 candidate models with different combina- In a second step, additional variables were included after
tions of the individual covariates was built (Supplementary testing for multi-collinearity. Variance inflation factors
material Appendix 1 Table A1). Similar to the procedure were below 2 and indicated no problems with collinearity
performed by Dinsmore et al. (2002) or Dinsmore and (Zuur et al. 2009). Thus, we added the variables number of
Dinsmore (2007), we used a hierarchical-modeling approach nestlings (on day 1, 5, 10 for the corresponding period), alti-
to keep the model set small, since the number of covariates tude (of the nest box), Julian day (clutch completely hatched,
was large. First, we performed a set of six models containing day 1 = 1 Apr) and average daily minimum ambient tem-
a constant, linear and quadratic time trend for daily survival perature (for the same 5-day period as for the precipitation
rate within and between years (Model 1–6). These models variable) to the best model of step 1 (global model). We did
should reflect time dependent survival rates (Grant et al. not include an interaction term between temperature and
2005, Berkunsky et al. 2016). Then, we used the model which precipitation, because Pearson’s correlation coefficient never
explained temporal variations within and between years best, exceeded 0.3. Since predictor variables had very different
and added the individual covariates altitude (Slobodník et al. scales, we rescaled them by generating standardized scores
2013), nest age (Grant et al. 2005) and number of hatchlings (z-score, subtracting the variable mean and dividing by the
(Mertens 1969) separately and additively (Model 7–10). variable standard deviation) to avoid problems with model
Finally, we included the weather covariates by adding each convergence. Then, this global model was used to generate a
one separately and also the interaction terms of temperature set of models with all possible combinations of fixed effects
and precipitation (Model 11–13) to the best model from the (R package ‘MuMIn’, Bartoń 2018). The models were graded
set of models we had run before. Models were ranked using according to their AICc, and model-averaged coefficients for
Akaike’s information criterion corrected for small sample size a subset of models (ΔAICc < 2) were extracted. Since Akaike
(AICC) to determine the best model (Burnham and Anderson weights of the best model was below 0.9 (Results, Table 3)
2002). and therefore high model selection uncertainty existed, full-
model averaging was used (Grueber et al. 2011, Symonds and
Modeling of nestling survival Moussalli 2011). Upper and lower bounds of the 95% confi-
dence intervals were calculated for each parameter. Statistical
To test if weather conditions affect not only nest sur- analyses were conducted using the software R 3.4.1 (< www.r-
vival during the nestling period, but also survival of project.org >).

4
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Weather conditions measured at the meteorological station between 22 Apr and 19 June and the percentage of successful nests in
the six study years. 30-year average (1981–2010) of precipitation measured between April and June exhibited 231 ml.
Ambient temperature (°C) Daily precipitation (mm) Successful nests
Study year Min Max Mean (± SE) Min Max Mean (± SE) Sum (%)
2010 0.4 29.0 12.1 (0.08) 0 40.7 6.1 (1.27) 362.53 56.8
2011 −1.3 24.1 13.2 (0.07) 0 45.6 3.7 (1.21) 217.50 80.4
2012 0.7 27.0 13.4 (0.09) 0 21.4 2.7 (0.63) 160.50 84.5
2013 1.7 28.5 12.6 (0.08) 0 47.9 4.9 (1.28) 288.00 67.4
2014 0.3 29.3 12.7 (0.08) 0 98.6 5.3 (1.88) 313.10 63.6
2015 2.1 27.8 13.2 (0.08) 0 60.0 3.8 (1.33) 222.40 69.0

Results Modeling daily nest survival

Breeding data Within the set of six preliminary candidate models, which
included constant, linear and quadratic time trends and
Within the six study years, out of the 350 nests, 252 nests therefore allowed for temporal variation within and between
were successful and 98 nests failed before nestlings were ready years, the model including a linear time trend and study year
to fledge. Nearly 42% of nest failure was due to abandon- received stronger support than the other models. Adding the
ment and 58% to nest predation. Compared to the other individual covariates altitude, nest age or number of nestlings
study years, in 2012 a high number of the provided nest when the nest was first checked during the nestling period
boxes was occupied by great tits; almost twice as many nests did not substantially improve the model fit. However, the fit
as monitored in the other five study years (Supplementary of the model was improved when we added weather condi-
material Appendix 1 Table A2). In total, 1640 nestlings sur- tions to this preselected model (Table 2).
vived until day 15 after hatch. Mean number of hatchlings The final best model including a linear time trend, study
per nest was lowest in the year 2012 (Kruskal–Wallis rank year and daily sum of precipitation [confidence intervals
sum test, Chi-squared = 49.42, df = 5, p < 0.001, followed CI = −0.047, −0.023] carried the most support (wi = 65%,
by Wilcoxon rank sum tests with Bonferroni adjustment, Table 2). Daily sum of precipitation had a negative effect on
p < 0.05, see Supplementary material Appendix 1 Table A2), nest survival (βP = −0.04 [−0.047, −0.023], Fig. 1), whereas
whereas the number of hatchlings per nest did not differ in the confidence interval for the effect of the linear time trend
the other years (Wilcoxon rank sum tests with Bonferroni included zero (βL = −0.02 [−0.057, 0.010]). We also ran a
adjustment, p > 0.05). Although brood size (mean number post-hoc model including both a linear and a quadratic
of nestlings) was low in the year 2012, the high occupancy effect of rainfall (βP = −0.03 [−0.077, −0.001], βPP = 0.00004
rate (almost 50% of nest boxes were occupied) and high over- [−0.0004, 0.0005]), since weather conditions could have a
all nest survival (85% successful nests) led to a high num- non-linear influence on nest survival (Mattsson and Cooper
ber of fledglings in the year 2012 (Supplementary material 2009, Post van der Burg et al. 2010), but it did not improve
Appendix 1 Table A2). the model fit (AICc = 688.80).

Table 2. Model selection table for great tit nest survival in the six
Weather conditions
study years ranked by ascending ΔAICc using the software MARK.
Variables included: linear time trend (L), quadratic time trend (Q),
Weather conditions in the nestling period between 22-Apr constant time trend (null model, .), nest age, altitude (A), number of
and 19-June varied within and between years (Table 1). hatchlings (NN), daily sum of precipitation (P), minimum daily tem-
Daily sum of precipitation varied between 0 and 98.6 ml perature (T).
(mean ± SE = 4.4 ± 0.54 ml) and ambient temperature
recorded every 15 min at the weather station through- Model Deviance K AICC ΔAICc wi
out the nestling period varied between −1.3 and 29.3°C SL +Y + P 670.80 8 686.83 0.00 0.65
(mean ± SE = 12.9 ± 0.03) in the study period. Average sum SL+Y+P×T 668.03 10 688.08 1.25 0.35
of precipitation measured between April and June in the SL+Y+T 685.24 8 701.27 14.44 0.00
SL+Y 689.63 7 703.66 16.83 0.00
30-year period between 1981 and 2010 exhibited 231 ml. SL+Y+NEST AGE 688.16 8 704.19 17.36 0.00
Compared to this prolonged period, we conducted our study SL+Y+A 689.26 8 705.29 18.46 0.00
during two wet springs (years 2010, 2014), one dry spring SL+Y+NN 689.46 8 705.49 18.66 0.00
(year 2012) and three intermediate springs (years 2011, SQ+Y 692.61 7 706.64 19.81 0.00
2013, 2015). The rainiest and coldest spring in our study S(.)+Y 695.42 6 707.44 20.61 0.00
period was spring 2010. In this year, only slightly more than SL 703.93 2 707.93 21.10 0.00
half of the nests were successful. In contrast, spring 2012 SL+Y+A+NEST AGE+NN 687.90 10 707.95 21.12 0.00
was the warmest and driest spring and nest survival reached SQ 706.26 2 710.26 23.43 0.00
S(.) 711.01 1 713.01 26.18 0.00
almost 85% (Table 1).

5
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
number of nestlings. Increasing number of days with heavy
rainfall negatively affected nestling survival. Confidence
intervals for the parameters Julian days, temperature, altitude
and number of nestlings indicated little evidence of affecting
nestling survival (Table 4).
For modeling nestling survival between day 11 and 15
(n = 276) the subset encompassed five models (Table 3). All
final models describing variation in nestling survival included
the variables number of days with heavy rainfall (> 20 mm)
and Julian days. In addition, some models included the vari-
ables average daily minimum temperature, altitude and num-
ber of nestlings on day 10 (Table 4). Number of days with
heavy precipitation and Julian days (after clutches completely
hatched) negatively affected nestling survival. The variables
average daily minimum temperature, altitude and number of
nestlings had no effect on nestling survival between day 11
and 15 (confidence intervals included zero).

Figure 1. Effect of daily sum of precipitation (mm) on daily nest


survival rate (± 95% CI) of great tits in Austria (2010–2015). For Discussion
computing estimates of real parameters, the effect of the linear time
trend was fixed (Table 2). Each point in the figure refers to a single Our results indicated that local weather conditions explained
great tit nest (n = 350). most variation in daily survival rate of great tit nestlings; in
particular precipitation was a dominant driver of reproduc-
One additional model shared a substantial weight of evi- tive success. Precipitation led to brood reduction through
dence in the confidence set, a model including next to the selective individual nestling death, but it ultimately also had
linear time trend and study year also an interaction between a negative effect on the entire brood. Further variables, which
precipitation and daily minimum ambient temperature mea- were included in the analyses of nest survival (altitude, nest
sured over 24 h (wi = 35%). However, the confidence interval age, number of nestlings) did not explain variation in nest
for the effect of the interaction term of precipitation and tem- fate. Clear evidence for the effect of ambient temperature
perature included zero (model SL+Y+P×T: βP×T = 0.005 [−0.002, on overall nest fates or survival of individual nestlings could
0.012]). Annual average DNS during the nestling period not be found (Bordjan and Tome 2014), but birds seem to
varied between 0.975 and 0.984 and apparent nestling sur- respond to year-to-year changes in climatic conditions in an
vival within the 15-day period between hatching and fledging ecological time frame (Skagen and Adams 2012); and as a
ranged between 57.9 and 78.9% (Supplementary material consequence the probability that nestlings survive the period
Appendix 1 Fig. A1). between hatching and fledging varied in the six study years.
Analyses of nest fates using the program MARK and tradi-
Modeling nestling survival tional approaches analysing nestling survival using GLMMs
showed similar results, thereby showing that both analyses
For modeling early nestling survival (day 1–5, n = 348), the can complement and support each other. Thus, analysing sur-
subset of models with ΔAICc < 2 encompassed nine mod- vival on a coarse level can already provide good insights in
els (Table 3). The final model describing nestling survival the complex interactions and processes of how environmen-
between day 1 and 5 included the variables number of heavy tal conditions affect nest survival. The nest survival analysis
rain days (> 20 mm), number of nestlings, average daily min- using MARK is highly advantageous especially for studies in
imum temperature, altitude and Julian days. Nestling sur- natural systems, where survival is not checked on a regular
vival within the first five days after clutches were completely basis. We assume that precipitation caused variation in indi-
hatched decreased with increasing number of days with heavy vidual nestling survival and nest fate: 1) directly, as it might
rainfall (> 20 mm). Confidence intervals for the parameter have affected thermoregulation and 2) indirectly, through its
estimates of number of nestlings, temperature, altitude and effect on food availability (Schöll et al. 2016) and predation
Julian days included zero, indicating little evidence that these risk. However, while reduced food availability most likely
variables affected nestling survival within the first 5 days after might have led to brood reduction through selective individ-
hatching (Table 4). ual nestling death (Siikamäki 1996), predation might have
The subset of models explaining variation in nestling sur- caused the loss of the entire brood leading to nest failures.
vival between day 6 and 10 (n = 317) encompassed six mod- Nestlings cannot maintain their own body temperature
els (Table 3). The final models with ΔAICc < 2 included the during the first days after hatch (Mertens 1977, Whittow
parameters number of heavy rain days (> 20 mm), Julian 2000). Feathers are not yet fully developed and nestlings
days, average daily minimum temperature, altitude and easily suffer immediate physical damage during this period

6
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 3. Subsets of generalized linear mixed effect models (ΔAICc < 2) explaining the variation in nestling survival in great tits based on a
multi-model inference approach. Variables: number of heavy rain days (n heavy rain), number of nestlings (n nestlings), average daily mini-
mum temperature (temp), Julian days clutch completely hatched (J days), altitude.
Response Model subsets AICc ΔAICc AICc weight
Nestling survival day 1–5 = Intercept + n heavy rain + n nestlings 614.4 0.0 0.197
= Intercept + n heavy rain 614.8 0.4 0.167
= Intercept + n heavy rain + n nestlings + temp 615.2 0.8 0.133
= Intercept + n heavy rain + n nestlings + J days 615.5 1.1 0.117
= Intercept + n heavy rain + temp 616.2 1.8 0.080
= Intercept + n heavy rain + n nestlings + altitude 616.3 1.9 0.079
= Intercept 616.4 2.0 0.076
= Intercept + n heavy rain + n nestlings + temp + J days 616.4 2.0 0.075
= Intercept + n heavy rain + J days 616.4 2.0 0.075
Nestling survival day 6–10 = Intercept + n heavy rain + J days 669.6 0.0 0.304
= Intercept + n heavy rain + J days + temp 670.7 1.1 0.180
= Intercept + n heavy rain 671.1 1.5 0.147
= Intercept + n heavy rain + J days + altitude 671.3 1.7 0.130
= Intercept + n heavy rain + J days + n nestlings 671.5 1.9 0.121
= Intercept + n heavy rain + temp 671.5 1.9 0.119
Nestling survival day 11–15 = Intercept + n heavy rain + J days 742.3 0.0 0.318
= Intercept + n heavy rain + J days + temp 743.0 0.7 0.220
= Intercept + n heavy rain + J days + altitude 743.1 0.8 0.214
= Intercept + n heavy rain + J days + n nestlings 744.1 1.8 0.130
= Intercept + n heavy rain + J days + temp + altitude 744.2 1.9 0.119

(Newton 1998). To reduce heat loss during cold and rainy provisioning behavior (Radford et al. 2001, Öberg et al. 2015)
periods, nestlings can clump and roost together, thus, brood and might face a trade-off between brooding nestlings and
size can affect energy expenditure for heat production and feeding themselves (Bordjan and Tome 2014). Ultimately,
therefore nestling survival (Mertens 1969). However, survival adult birds might be forced to abandon nests (Bordjan and
of individual nestlings was not related to the number of sib- Tome 2014) to cover their own energy demands, if meta-
lings in our study site. Temporal patterns (e.g. nestling age) bolic costs of searching for food for nestlings are too high.
seem to affect nest survival (Grant et al. 2005, Boulton et al. Within our long-term study, 11.7% of all nests were aban-
2008), because shortly after hatching young nestlings are doned between 2010 and 2015. When nestlings grow older,
incapable of thermoregulation (Whittow 2000) and rely on gain body weight and change to an endothermic metabolism
females to keep them warm (Kluijver 1951). If females spent (Mertens 1977), females can slowly reduce their brooding
more time within nest boxes, they have to reduce their food behavior and increase time spent searching for food (Kluijver

Table 4. Model-averaged summary statistics of generalized linear mixed effect models: estimates of coefficients and unconditional standard
error (SE), lower and upper limits of the 95% confidence intervals (CI 95%) and relative importance values of the model parameters.
Important parameters affecting response variables are highlighted in bold. Variables: number of heavy rain days (n heavy rain), number of
nestlings (n nestlings), average daily minimum temperature (temp), Julian days clutch completely hatched (J days), altitude.
Response Variables Estimate SE CI 95% Relative importance
Nestling survival day 1–5 (Intercept) 5.536 0.607 4.347; 6.725
n heavy rain −0.432 0.219 −0.861; −0.004 1.00
n nestlings 0.373 0.223 −0.065; 0.811 0.60
temp 0.232 0.229 −0.218; 0.681 0.29
altitude −0.106 0.214 −0.526; 0.314 0.27
J days −0.207 0.232 −0.662; 0.248 0.08
Nestling survival day 6–10 (Intercept) 4.768 0.466 3.855; 5.681
n heavy rain −0.616 0.205 −1.017; −0.214 1.00
J days −0.393 0.218 −0.820; 0.033 0.73
temp −0.244 0.220 −0.676; 0.188 0.30
altitude 0.132 0.214 −0.288; 0.552 0.13
n nestlings −0.109 0.224 −0.549; 0.330 0.12
Nestling survival day 11–15 (Intercept) 2.943 0.196 2.558; 3.327
n heavy rain −0.535 0.136 −0.801; −0.268 1.00
J days −0.359 0.155 −0.663; −0.055 1.00
temp 0.179 0.165 −0.145; 0.502 0.34
altitude 0.155 0.147 −0.134; 0.443 0.33
n nestlings 0.079 0.145 −0.204; 0.363 0.13

7
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1951). Adult birds and nestlings also suffer large heat loss if nest failure and therefore, analyses of nest survival should also
they are exposed to rainfall or get wet. Short periods and low include predated nests (but see Bordjan and Tome 2014).
intensity of rain might be less problematic, since plumage of Since food availability is affected by weather conditions and
adult birds is at least temporarily rainproof (Kennedy 1970) reduced during periods with harsh weather (Schöll et al.
and nest boxes can shield drizzle. However, heavy precipita- 2016), an increased begging behavior of hungry nestlings
tion events might have negatively affected nestling survival, (Kölliker et al. 1998) might have attracted predators (Leech
if the nesting material got soaked (Wesolowski et al. 2002). and Leonard 1997, Fogarty et al. 2017). In addition, as nest-
In addition, if adult birds released moisture within nest boxes ing season advances and nestlings mature, nestlings require
during food provisioning, nestlings can rapidly lose body more food (García-Navas and Sanz 2011). Since parental
mass (Whitehouse et al. 2013) during rainy periods, because activity at the nest increases, because more frequent nest
energy demands to maintain body temperature constant visits are necessary to provide food for nestlings, the prob-
increase if birds’ plumage is wet (reviewed by Kennedy 1970, ability of predation due to diurnal predators increases as well
Wilson et al. 2004). (Martin et al. 2000). Indeed, predation occurred in 16.3%
Nestlings of many bird species and great tits are mainly of all studied nests (n = 350) throughout the 15-day nestling
fed with invertebrate prey (Naef-Daenzer and Keller 1999, period in our study site. In contrast to studies from Lithuania
Naef-Daenzer et al. 2000). Caterpillars, e.g. of winter moth (Juškaitis 2006), Germany (Koppmann-Rumpf et al. 2003)
Operophtera brumata, are one of the main food sources of or Czech Republic (Adamík and Král 2008), where only
great tit nestlings, especially in study sites, where they occur 9.6%, 8.4% and respectively 3.2% of great tit nests were
only during a short annual peak (Naef-Daenzer and Keller destroyed in the prolonged period between egg-laying and
1999). In our study site, caterpillars are available throughout fledging, we showed that predation had a huge effect on nest
the great tit breeding season (Schöll et al. 2016), but nestling survival. However, compared to a study of great tits nesting in
survival between day 11 and 15 after hatch was negatively natural tree cavities in Poland, where predation accounted for
associated with the Julian day of hatching (clutch completely 66.7% of nest failures (Maziarz et al. 2015), predation risk
hatched, day 1 = 1 Apr). was lower for our great tit population breeding in nest boxes
Since insect development itself is highly dependent on (accounted for 58% of nest failures in our study site, see also
temperature and rainfall (Price 1997), weather conditions Purcell et al. 1997, but see Mitrus 2003).
should also have an indirect effect on individual nestling One of the most important predators in our study site,
survival. During rainy periods, caterpillars are less active the edible dormouse Glis glis, is mainly active between sunset
(Tamburini et al. 2013) and might be difficult to detect by and sunrise (Rodolfi 1994). It does not only depredate eggs
great tits. Caterpillar availability is reduced during harsh and nestlings; it can also predate adult birds resting within
weather conditions (Schöll et al. 2016), because heavy and nest boxes (Adamík and Král 2008). Until now, it is not clear,
persistent rainfall increases mortality of caterpillars (Dennis whether dormice find occupied nest boxes just by chance due
and Sparks 2007). As food availability is one factor limit- to a overlap in timing of breeding seasons, or whether they
ing nestling survival (Boulton et al. 2008), optimal weather actively search for occupied nest boxes (Adamík and Král
conditions in the year 2012 might have been advantageous 2008). Although great tits might be able to defend cavities
for caterpillar survival and development (Newton 1998). against intruders (Adamík and Král 2008), they might prob-
In contrast, low food availability during years with harsh ably abandon nests after a dormouse tried to enter, because
weather conditions might have negatively affected nestling of high predation risk. It is assumed, that local abundance
survival, because parents could not provide enough food for of predators increases as the breeding season progresses and
the whole brood, or might have been forced to abandon nests therefore, nest survival declines throughout the nesting sea-
to increase their own chances of survival. Moreover, adult son (Grant et al. 2005). However, during the last decades,
birds seek shelter during strong rainfall events and reduce activity seasons of the edible dormouse might have shifted
feeding rates (Radford et al. 2001, Öberg et al. 2015, but see into early spring and is assumed to overlap with the nesting
Geiser et al. 2008). As adult birds might not completely com- season of great tits (Koppmann-Rumpf et al. 2003, but see
pensate for periods of rainfall and lack of food by increasing Bieber et al. 2018).
feeding frequency after rainfall has stopped (Radford et al. Our study showed that local weather conditions in general,
2001), individual nestlings can either die from hypothermia and heavy and persistent rainfall in particular had a strong
or starvation, if they do not receive enough food to meet the negative effect on nestling survival. Among all years, mean
demands of heat production, which increase during low tem- number of hatchlings per nest box was the lowest in spring
perature periods (Newton 1998). 2012. In this year, almost 50% of the available nest boxes were
Bordjan and Tome (2014) have also studied the effects of occupied, and the high nest box breeding density might have
weather conditions on nest survival of great tits. They found led to high intraspecific competition. Thus, a strong density-
indications that rainfall might cause nest abandonment, and dependent effect might have resulted in the low average num-
reasoned that food-shortages during cold and wet periods had ber of hatchlings in the year 2012 (Van Balen 1973, Dhondt
affected nest fate. However, in many passerine species, nest 2010). However, in contrast to the other study years, the
predation is one of the most important reasons for complete high overall nest survival (85% successful nests) during the

8
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
nestling period finally resulted in a high number of fledglings Bartoń, K. 2018. MuMIn: multi-model inference. R package ver.
in this year. Spring 2012 was the driest spring in our study 1.40.4, <https://cran.r-project.org/web/packages/MuMIn/
site and these conditions seemed to increase the probability index.html>.
of nestling survival until fledging, compared to wet years, Bates, D., Maechler, M., Bolker, B., Walker, S., Christensen, R. H.
where number of hatchlings was higher, but nest survival was B., Singmann, H. and Dai, B. 2014. lme4: linear mixed-effects
models using Eigen and S4. R package ver. 1.1-7. – <https://
reduced and finally fewer nestlings had fledged. In contrast cran.r-project.org/web/packages/lme4/index.html>.
to our study, extremely harsh weather conditions occurred Berkunsky, I., Segura, L. N., Aramburú, R. M., Ruggera, R. A.,
during spring 2012 in eastern England and affected breeding Svagelj, W. S. and Reboreda, J. C. 2016. Nest survival and
success negatively (Whitehouse et al. 2013). Thus, studying predation in blue-fronted parrots Amazona aestiva: effects of
geographic variation of weather patterns seems important to nesting behaviour and cavity characteristics. – Ardea 104:
model animal population trends and will be of high ecological 143–151.
and conservation interest, since the Intergovernmental Panel Bieber, C., Turbill, C. and Ruf, T. 2018. Effects of aging on timing
on Climate Change (IPCC) forecasts an increasing number of hibernation and reproduction. – Sci. Rep. 8: 13881.
of heavy rain events in future (IPCC 2013). Ambient tem- Bordjan, D. and Tome, D. 2014. Rain may have more influence
perature and precipitation do not only affect survival of nest- than temperature on nest abandonment in the great tit Parus
major. – Ardea 102: 79–85.
lings, they can have a complex interacting impact on different
Both, C. and Visser, M. E. 2005. The effect of climate change on
age classes, e.g. recruitment of fledglings (Öberg et al. 2015, the correlation between avian life-history traits. – Global
Gyurácz et al. 2016) or parental survival (Öberg et al. 2015). Change Biol. 11: 1606–1613.
Both, C., Artemyev, A. V., Blaauw, B., Cowie, R. J., Dekhuijzen, A.
J., Eeva, T., Enemar, A., Gustafsson, L., Ivankina, E. V., Järvinen,
Acknowledgements – Thanks to M. Polo Aparisi, M. de los Milagros
A., Metcalfe, N. B., Nyholm, N. E. I., Potti, J., Ravussin, P. A.,
Alvarez Martinez and Raimund Barth for help with fieldwork.
Sanz, J. J., Silverin, B., Slater, F. M., Sokolov, L. V., Török, J.,
Further, we thank two anonymous reviewers for their helpful
Winkel, W., Wright, J., Zang, H. and Visser, M. E. 2004. Large-
comments on the original draft of the manuscript.
scale geographical variation confirms that climate change causes
Funding – This study was funded by the Federal Ministry for
birds to lay earlier. – Proc. R. Soc. B 271: 1657–1662.
Sustainability and Tourism, Austria; Wolfgang Pauli Fellowship,
Boulton, R. L., Richard, Y. and Armstrong, D. P. 2008. Influence
Univ. of Vienna and the Austrian Federal Forests.
of food availability, predator density and forest fragmentation
Author contributions – The first author collected data in the field,
on nest survival of New Zealand robins. – Biol. Conserv. 141:
analyzed the data, performed statistical analysis and wrote the
580–589.
manuscript. The second author originally developed methodology
Burnham, K. P. and Anderson, D. R. 2002. Model selection and
and supervised the project.
multimodel inference: a practical information–theoretic
Conflicts of interest – There is no conflict of interest.
approach. – Springer-Verlag.
Permit(s) – Permission to handle our study animals was given by the
Dennis, R. L. H. and Sparks, T. H. 2007. Climate signals are
Office of the Provincial Government of Lower Austria. The Austrian
reflected in an 89 year series of British Lepidoptera records.
Federal Forests gave their permission to work on their land.
– Eur. J. Entomol. 104: 763–767.
Dinsmore, S. J. and Dinsmore, J. J. 2007. Modeling avian nest
surival in program MARK. – Stud. Avian Biol. 34: 73–83.
Transparent Peer Review Dinsmore, S. J., White, G. C. and Knopf, F. L. 2002. Advanced
techniques for modeling avian nest survival. – Ecology 83:
The peer review history for this article is available at https:// 3476–3488.
publons.com/publon/10.1111/jav.02418 Dhondt, A. A. 2010. Effects of competition on great and blue tit
reproduction: intensity and importance in relation to habitat
quality. – J. Anim. Ecol. 79: 257–265.
References Fogarty, D. T., Elmore, R. D., Fuhlendorf, S. D. and Loss, S. R.
2017. Influence of olfactory and visual cover on nest site selec-
Adamík, P. and Král, M. 2008. Nest losses of cavity nesting birds tion and nest success for grassland-nesting birds. – Ecol. Evol.
caused by dormice (Gliridae, Rodentia). – Acta Theriol. 53: 7: 6247–6258.
185–192. García-Navas, V. and Sanz, J. J. 2011. The importance of a main
Arlt, D., Forslund, P., Jeppsson, T. and Part, T. 2008. Habitat- dish: nestling diet and foraging behaviour in Mediterranean
specific population growth of a farmland bird. – PLoS One 3: blue tits in relation to prey phenology. – Oecologia 165:
e3006. doi: 10.1371/journal.pone.0003006 639–649.
Armstrong, D. P., Raeburn, E. H., Powlesland, R. G., Howard, M., Geiser, S., Arlettaz, R. and Schaub, M. 2008. Impact of weather
Christensen, B. and Ewen, J. G. 2002. Obtaining meaningful variation on feeding behaviour, nestling growth and brood sur-
comparisons of nest success: data from New Zealand robin (Pet- vival in wrynecks Jynx torquilla. – J. Ornithol. 149: 597–606.
roica australis) populations. – N. Z. J. Ecol. 26: 1–13. Grant, T. A., Shaffer, T. L., Madden, E. M. and Pietz, P. J. 2005.
Auer, I., Chimani, B. and Türk, K. 2012. Klimanormalperiode Time-specific variation in passerine nest survival: new insights
1981–2010. – In: Geodynamik, Z. F. M. U. (ed.), 13. Öster- into old questions. – Auk 122: 661–672.
reichischer Klimatag, 14.02.2012–15.02.2012, Wien. Grueber, C. E., Nakagawa, S., Laws, R. J. and Jamieson, I. G. 2011.
Avery, M. I. and Krebs, J. R. 1984. Temperature and foraging success Multimodel inference in ecology and evolution: challenges and
of great tits Parus major hunting for spiders – Ibis 126: 33–38. solutions. – J. Evol. Biol. 24: 699–711.

9
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Gyurácz, J., Bánhidi, P., Góczán, J., Illés, P., Kalmár, S., Lukács, Mertens, J. a. L. 1969. The influence of brood size on the energy
Z., Németh, C. and Varga, L. 2016. Temperature and precipi- metabolism and water loss of nestling great tits Parus major
tation effects on breeding productivity of some passerines major. – Ibis 111: 11–16.
– a multivariate analysis of constant effort mist-netting data. Mertens, J. a. L. 1977. Thermal conditions for successful breeding
– Biologia 71: 1298–1303. in great tits (Parus major L.) – I. Relation of growth and devel-
IPCC 2013. Summary for policymakers. – In: Stocker, T. F., Qin, opment of temperature regulation in nestling great tits. – Oec-
D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nau- ologia 28: 1–29.
els, A., Xia, Y., Bex, V. and Midgley, P. M. (eds), Climate Change Mitrus, C. 2003. A comparison of the breeding ecology of collared
2013: the physical science basis. Contribution of Working flycatchers nesting in boxes and natural cavities. – J. Field Orni-
Group I to the Fifth Assessment Report of the Intergovernmen- thol. 74: 293–299.
tal Panel on Climate Change. Cambridge Univ. Press, pp. 3–29. Morrison, S. A. and Bolger, D. T. 2002. Variation in a sparrow’s
Jehle, G., Yackel Adams, A. A., Savidge, J. A. and Skagen, S. K. reproductive success with rainfall: food and predator-mediated
2004. Nest survival estimation: a review of alternatives to the processes. – Oecologia 133: 315–324.
Mayfield estimator. – Condor 106: 472–484. Naef-Daenzer, B. and Keller, L. F. 1999. The foraging performance
Juškaitis, R. 2006. Interactions between dormice (Gliridae) and of great and blue tits (Parus major and P. caeruleus) in relation
hole-nesting birds in nestboxes. – Folia Zool. 55: 225–236. to caterpillar development, and its consequences for nestling
Keller, L. F. and Van Noordwijk, A. J. 1994. Effects of local envi- growth and fledging weight. – J. Anim. Ecol. 68: 708–718.
ronmental conditions on nestling growth in the great tit Parus Naef-Daenzer, L., Naef-Daenzer, B. and Nager, R. G. 2000. Prey
major L. – Ardea 82: 349–362. selection and foraging performance of breeding great tits Parus
Kennedy, R. J. 1970. Direct effects of rain on birds: a review. – Brit. major in relation to food availability. – J. Avian Biol. 31:
Birds 63: 401–414. 206–214.
Kluijver, H. N. 1951. The population ecology of the great tit, Parus Newton, I. 1998. Population limitation in birds. – Academic Press
m. major L. – Ardea 39: 1–135. Limited.
Knape, J. and De Valpine, P. 2011. Effects of weather and climate O’Connor, R. J. and Morgan, R. A. 1982. Some effects of weather
on the dynamics of animal population time series. – Proc. Biol. conditions on the breeding of the spotted flycatcher Muscicapa
Sci. 278: 985–92. striata in Britain. – Bird Study 29: 41–48.
Kölliker, M., Richner, H., Werner, I. and Heeb, P. 1998. Begging Öberg, M., Arlt, D., Pärt, T., Laugen, A. T., Eggers, S. and Low,
signals and biparental care: nestling choice between parental M. 2015. Rainfall during parental care reduces reproductive
feeding locations. – Anim. Behav. 55: 215–222. and survival components of fitness in a passerine bird. – Ecol.
Koppmann-Rumpf, B., Heberer, C. and Schmidt, K.-H. 2003. Long Evol. 5: 345–356.
term study of the reaction of the edible dormouse Glis glis (Roden- Post van der Burg, M., Powell, L. A. and Tyre, A. J. 2010. Finding
tia: Gliridae) to climatic changes and its interactions with hole- the smoothest path to success: model complexity and the con-
breeding passerines. – Acta Zool. Acad. Sci. Hung. 49: 69–76. sideration of nonlinear patterns in nest-survival data. – Condor
Laake, J. 2018. Package ‘RMark’. R package ver. 2.2.5. – <https:// 112: 421–431.
cran.r-project.org/web/packages/RMark/index.html>. Price, P. W. 1997. Insect ecology. – Wiley.
Leech, S. M. and Leonard, M. L. 1997. Begging and the risk of Purcell, K. L., Verner, J. and Oring, L. W. 1997. A comparison of
predation in nestling birds. – Behav. Ecol. 8: 644–646. the breeding ecology of birds nesting in boxes and tree cavities.
Mainwaring, M. C., Hartley, I. R., Lambrechts, M. M. and Deem- – Auk 114: 646–656.
ing, D. C. 2014. The design and function of birds’ nests. – Ecol. Radford, A. N., Mccleery, R. H., Woodburn, R. J. W. and More-
Evol. 4: 3909–28. croft, M. D. 2001. Activity patterns of parent great tits Parus
Martin, K., Wilson, S., Macdonald, E. C., Camfield, A. F., Martin, major feeding their young during rainfall. – Bird Study 48:
M. and Trefry, S. A. 2017. Effects of severe weather on repro- 214–220.
duction for sympatric songbirds in an alpine environment: Ricklefs, R. E. 1969. An analysis of nesting mortality in birds.
interactions of climate extremes influence nesting success. – Smithsonian Contr. Zool. 9: 1–48.
– Auk 134: 696–709. Rose, J. P., Ersan, J. S. M., Wylie, G. D., Casazza, M. L. and Hal-
Martin, T. E., Scott, J. and Menge, C. 2000. Nest predation stead, B. J. 2019. Demographic factors affecting population
increases with parental activity: separating nest site and parental growth in giant gartersnakes. – J. Wildl. Manage. 83:
activity effects. – Proc. Biol. Sci. 267: 2287–2293. 1540–1551.
Matter, S. F. and Roland, J. 2017. Climate and extreme weather Rotella, J. 2018. Chapter 17. Nest survival models. – In: Cooch,
independently affect population growth, but neither is a con- E. and White, G. C. (eds), Program MARK: a gentle introduc-
sistently good predictor. – Ecosphere 8: e01816. tion, 18th edn. – <http://www.phidot.org/software/mark/docs/
Mattsson, B. J. and Cooper, R. J. 2009. Multiscale analysis of the book/>.
effects of rainfall extremes on reproduction by an obligate ripar- Rotella, J., Taper, M., Stephens, S. and Lindberg, M. 2007. Extend-
ian bird in urban and rural landscapes. – Auk 126: 64–76. ing methods for modeling heterogeneity in nest-survival data
Mayfield, H. 1961. Nesting success calculated from exposure. using generalized mixed models. – In: Jones, S. L. and Geupel,
– Wilson Bull. 73: 255–261. G. R. (eds), Beyond Mayfield: measurements of nest-survival
Mayfield, H. 1975. Suggestions for calculating nest success. – Wil- data. Cooper Ornithological Society, pp. 34–44.
son Bull. 87: 456–466. Rodolfi, G. 1994. Dormice Glis glis activity and hazelnut consump-
Maziarz, M., Wesołowski, T., Hebda, G., Cholewa, M. and tion. – Acta Theriol. 39: 215–220.
Broughton, R. K. 2015. Breeding success of the great tit Parus Schmidt, B. R., Feldmann, R. and Schaub, M. 2005. Demographic
major in relation to attributes of natural nest cavities in a pri- processes underlying population growth and decline in Sala-
meval forest. – J. Ornithol. 157: 343–354. mandra salamandra. – Conserv. Biol. 19: 1149–1156.

10
1600048x, 2020, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jav.02418 by Univ de Buenos Aires, Wiley Online Library on [03/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Schöll, E. M., Ohm, J., Hoffmann, K. F. and Hille, S. M. 2016. population dynamics of the pine processionary moth in the
Caterpillar biomass depends on temperature and precipitation, southern Alps. – Clim. Change 121: 701–712.
but does not affect bird reproduction. – Acta Oecol. 74: 28–36. Van Balen, J. H. 1973. A comparative study of the breeding ecology
Siikamäki, P. 1996. Nestling growth and mortality of pied flycatch- of the great tit Parus major in different habitats. – Ardea 61: 1–93.
ers Ficedula hypoleuca in relation to weather and breeding effort. Wesolowski, T., Czeszczewik, D., Rowiński, P. and Walankiewicz,
– Ibis 138: 471–478. W. 2002. Nest soaking in natural holes – a serious cause of
Skagen, S. K. and Adams, A. a. Y. 2012. Weather effects on avian breeding failure? – Ornis Fenn. 79: 132–138.
breeding performance and implications of climate change. White, G. C. and Burnham, K. P. 1999. Program MARK: survival
– Ecol. Appl. 22: 1131–1145. estimation from populations of marked animals. – Bird Study
Slobodník, R., Balážová, M., Jandzik, D. and Baláž, M. 2013. Local 46: S120–S139.
weather differently affects collared flycatcher reproduction at Whitehouse, M. J., Harrison, N. M., Mackenzie, J. and Hinsley,
different altitudes. – Cent. Eur. J. Biol. 8: 1145–1152. S. A. 2013. Preferred habitat of breeding birds may be compro-
Stearns, S. C. 1992. The evolution of life histories. – Oxford Univ. mised by climate change: unexpected effects of an exceptionally
Press. cold, wet spring. – PLoS One 8: e75536. doi: 10.1371/journal.
Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., pone.0075536
Boschung, J., Nauels, A., Xia, Y., Bex, V. and Midgley, P. M. Whittow, G. C. 2000. Sturkie’s avian physiology. – Academic Press.
2013. Climate Change 2013: the physical science basis. Con- Wilson, G. R., Cooper, S. J. and Gessaman, J. A. 2004. The effects
tribution of Working Group I to the Fifth Assessment Report of temperature and artificial rain on the metabolism of Ameri-
of the Intergovernmental Panel on Climate Change. – Cam- can kestrels (Falco sparverius). – Comp. Biochem. Phys. A 139:
bridge Univ. Press. 389–394.
Symonds, M. R. E. and Moussalli, A. 2011. A brief guide to model World Meteorological Organization 2017. WMO guidelines on the
selection, multimodel inference and model averaging in behav- calculation of climate normals. – World Meteorological Organ-
ioural ecology using Akaike’s information criterion. – Behav. ization.
Ecol. Sociobiol. 65: 13–21. Zuur, A. F., Ieno, E. N., Walker, N. J., Saveliev, A. A. and Smith,
Tamburini, G., Marini, L., Hellrigl, K., Salvadori, C. and Battisti, G. M. 2009. Mixed effects models and extensions in ecology
A. 2013. Effects of climate and density-dependent factors on with R. – Springer.

Supplementary material (available online as Appendix jav-


02418 at < www.avianbiology.org/appendix/jav-02418 >).
Appendix 1.

11

You might also like