ArticlePDF Available

Shifts in caterpillar biomass phenology due to climate change and its impact on the breeding biology of an insectivorous bird

Authors:

Abstract and Figures

Timing of reproduction has major fitness consequences, which can only be understood when the phenology of the food for the offspring is quantified. For insectivorous birds, like great tits (Parus major), synchronisation of their offspring needs and abundance of caterpillars is the main selection pressure. We measured caterpillar biomass over a 20-year period and showed that the annual peak date is correlated with temperatures from 8 March to 17 May. Laying dates also correlate with temperatures, but over an earlier period (16 March-20 April). However, as we would predict from a reliable cue used by birds to time their reproduction, also the food peak correlates with these temperatures. Moreover, the slopes of the phenology of the birds and caterpillar biomass, when regressed against the temperatures in this earlier period, do not differ. The major difference is that due to climate change, the relationship between the timing of the food peak and the temperatures over the 16 March-20 April period is changing, while this is not so for great tit laying dates. As a consequence, the synchrony between offspring needs and the caterpillar biomass has been disrupted in the recent warm decades. This may have severe consequences as we show that both the number of fledglings as well as their fledging weight is affected by this synchrony. We use the descriptive models for both the caterpillar biomass peak as for the great tit laying dates to predict shifts in caterpillar and bird phenology 2005-2100, using an IPCC climate scenario. The birds will start breeding earlier and this advancement is predicted to be at the same rate as the advancement of the food peak, and hence they will not reduce the amount of the current mistiming of about 10 days.
Content may be subject to copyright.
GLOBAL CHANGE ECOLOGY
Marcel E. Visser Æ Leonard J. M. Holleman
Phillip Gienapp
Shifts in caterpillar biomass phenology due to climate change
and its impact on the breeding biology of an insectivorous bird
Received: 6 June 2005 / Accepted: 2 November 2005 / Published online: 3 December 2005
Springer-Verlag 2005
Abstract Timing of reproduction has major fitness con-
sequences, which can only be understood when the
phenology of the food for the offspring is quantified. For
insectivorous birds, like great tits (Parus major), syn-
chronisation of their offspring needs and abundance of
caterpillars is the main selection pressure. We measured
caterpillar biomass over a 20-year period and showed
that the annual peak date is correlated with tempera-
tures from 8 March to 17 May. Laying dates also cor-
relate with temperatures, but over an earlier period (16
March 20 April). However, as we would predict from a
reliable cue used by birds to time their reproduction,
also the food peak correlates with these temperatures.
Moreover, the slopes of the phenology of the birds and
caterpillar biomass, when regressed against the temper-
atures in this earlier period, do not differ. The major
difference is that due to climate change, the relationship
between the timing of the food peak and the tempera-
tures over the 16 March 20 April period is changing,
while this is not so for great tit laying dates. As a con-
sequence, the synchrony between offspring needs and the
caterpillar biomass has been disrupted in the recent
warm decades. This may have severe consequences as we
show that both the number of fledglings as well as their
fledging weight is affected by this synchrony. We use the
descriptive models for both the caterpillar biomass peak
as for the great tit laying dates to predict shifts in cat-
erpillar and bird phenology 2005–2100, using an IPCC
climate scenario. The birds will start breeding earlier and
this advancement is predicted to be at the same rate as
the advancement of the food peak, and hence they will
not reduce the amount of the current mistiming of about
10 days.
Keywords Climate change Æ Fitness Æ Great tit Æ
Phenology Æ Timing of reproduction
Introduction
Timing of reproduction is a life-history trait with
important fitness consequences (e.g. Daan et al. 1990;
Klomp 1970; Perrins 1970; Verhulst et al. 1995) because
for many species there is a short time-window in the
annual cycle where conditions are sufficiently go od so
that they can successfully reproduce. This time window
is often set by the abundance of food necessary to raise
offspring (Lack 1950; Martin 198 7). As the seasonal
pattern in the environmental conditions varies from year
to year, also the optimal time window for reproduction
varies annually. In accordance with this, timing of
reproduction shows corresponding within-individual
year-to-year variation, caused by phenotypic plasticity.
To understand the selection acting on the timing of
reproduction we need to identify the environmental
factors, for example, prey phenology, which determine
the optimal time window for reproduction. Only then we
can understand why, in an ultimate sense, the animals
reproduce early in one year and late in another, and why
within a year late-reproducing animals do better or
worse than early-reproducing conspecifics. What is
more, it is the only avenue to a more predictive study on
the impact of climate change on the reproductive output
of animals.
There is ample evidence that climate change has led
to changes in the phenology of many species (e.g. Beebee
1995; Brown et al. 1999; Crick et al. 1997; Parmesan
et al. 1999; Penuelas et al. 2002). Parmesan and Yohe
(2003) summarized the literature and concluded that
there is ‘‘...‘very high confidence’ that climate change is
already affecting living systems.’’ It is however unclear
how we should interpret this. If a species has shifted
10 days in 20 years, is that positive (adapting to
changing conditions) or negative (climate change is
affecting living systems)? What we need is a yardstick:
Communicated by Bernhard Stadler
M. E. Visser (&) Æ L. J. M. Holleman Æ P. Gienapp
Netherlands Institute of Ecology (NIOO-KNAW),
P.O. Box 40, 6666 ZG Heteren, The Netherlands
E-mail: m.visser@nioo.knaw.nl
Tel.: +31-26-4791253
Fax: +31-26-4723227
Oecologia (2006) 147: 164–172
DOI 10.1007/s00442-005-0299-6
how much should a species shift given the change in
temperatures and other climate variables (Visser and
Both 2005; Visser et al. 2004)? The yardstick is of course
the shift in the moment in the year the environment is
most suitable for reproduction; if this moment has
shifted 10 days as well, the focus species is doing very
well, if this moment has shifted 20 da ys, or not at all,
then obviously climate change has a negative effect.
In many taxa the main variable determining repro-
ductive success is the abundance of prey items (e.g. Dias
and Blon del 1996; Durant et al. 2003; Pearce-Higgins
and Yalden 2004; van Noordwijk et al. 1995; Verboven
et al. 2001) affecting also the workload of the adults
(Thomas et al. 2001). Yet, there are only a limited
number of systems where data is collected on the annual
cycle of a species and on the temporal variation of its
prey abundance. We will focus on one of the more well-
known systems of great tits (Parus majo r ) relying on
caterpillars in oaks as the main food source for their
nestlings (e.g. Naef-Daenzer et al. 2000; Perrins 1991). In
this system, the caterpill ar prey appears after the trees
have thei r bud burst and disappears when the caterpil-
lars are fully grown and pupate in the soil and are thus
no longer available. Hence, there is only a short period
of ample food, and these insectivorous birds need to
synchronise their breeding in such a way that the time of
the maximal need of their offspring for food coincides
with the time of maximal food abundance.
The phenology of caterpillars in trees is generally
measured indirectly by collecting caterpillar droppings
with so called frass-nets placed beneath a tree (e.g. Fis-
chbacher et al. 1998; Verbo ven et al. 2001; Visser et al.
1998). The term frass stems from a mistranslation of the
German word Fraß meaning ‘amount eaten away’ (from
a certain food) rather than dropping. The biomass of the
caterpillars is calculated from the mass of the droppings
collect in the nets corrected for temperature (Tinbergen
and Dietz 1994). Although this method has been vali-
dated (Fischbacher et al. 1998; Liebhold and Elkinton
1988a, b; Zandt 1994) it may be possible that weather
conditions (wind, rain) and properties of the tree (height,
size) introduce systematic errors. To validate the estab-
lished relationship further and possibly to correct for the
mentioned effects we compared caterpillar biomas s cal-
culated from collected droppings and from direct branch
samples from trees.
We have reported on the impact of climate change on
the synchronisation of great tit laying dates and cater-
pillar abundance in previous publications (Visser et al.,
1998, 200 4) where we showed that caterpillar peak bio-
mass date has shifted forward over the past decades but
that there has been no advancement of laying dates of
great tits. Here, we will take our studies on this system a
step further and predict how forecasted climate change
will impact the pheno logy of both caterpillar biomass
and bird reproduction in the next 100 years.
Our objectives are to describe the caterpillar biomass
changes throughout a season, to const ruct a descriptive
model of the peak date and width of the biomass curve.
We will also compare the temperatu re sensitivity of the
great tit laying dates and of the food peak to asses the
impact of changing tem peratures on the synchrony of
the birds with their main food. Furthermore, we corre-
late a number of great tit fitness parameters to the bio-
mass distribution to demonstrate the importance of this
synchronisation. Finally, we use a descriptive model to
predict the future caterpillar biomass distributions and
laying dates under IPCC scenarios and use these pre-
dicted caterpillar and bird phenologies to estimate the
impact of the changing food phenology on the birds’
breeding success.
Methods
Fieldwork
Fieldwork was carried out in the Hoge Veluwe, one of
our long-term (1955–2004) study areas. This study area
consists of 171 ha mixed woodland on poor sandy soils,
dominated by oak and pine with about 400 nest boxes.
Some additional comparisons are made with data from
another long-term study area, Oosterhout, a smaller,
isolated woodland (11 ha ) on rich clay soils near the
river Waal, in which 120 nest boxes are placed.
From 1993 to 2004 frass nets were placed under a
varying number of trees on the Hoge Veluwe, but there
is a fixed set of seven oak trees, located at different sites
within the study area (>500 m apart), which were
sampled throughout these years. Furthermore, from
1985 to 1992 frass nets were used in the same area by J.
M. Tinbergen, and we use these data on the phenology
of the caterpillar biomass (see Verboven et al. 2001)to
obtain a 20-year time series. For Oosterhout frass net
data are available for the years 1957–1967 (van Balen
1973), 1986–1988 (Zandt 1994), 1996 and 2001–2004
(own data).
Frass nets (a cheese cloth of 0.25 m
2
in a metal frame,
with a weight hung from the centre of the net, see Fig. 3a
in Tinbergen 1960) were put up under oak trees (Quercus
robur). Two of these nets are placed under a tree (about
1–1.5 m from the stem) and every 3–4 days all caterpil-
lar droppings are collected, dried at 60 C for 24 h,
sorted (i.e. all debris is removed), weighted and from this
the caterpillar biomass is calculated (see below).
During 3 years (1994–1996) we took branch samples
using a sky lift at the Hog e Veluwe. Branches of about
2 m were first put into a large plastic bag, then cut from
the trees and later (within 36 h) sorted for insects. The
length of the branches were measured to convert the
total biomass to biomass/meter branch. We also mea-
sured height and stem diameter of these trees.
To analyse the spatial variation in the time of
maximal caterpillar biomass we correlated the peak
date of caterpillar biomass with bud burst date of the
tree under which the nets were placed. Tree leaf phe-
nology was scored twice a week and we use the date
at which the leaves protruded from their buds in the
165
canopy of the tree as their bud burst date (Visser and
Holleman 2001).
Data on laying dates and clutch sizes of great tits in
the various populations was collected by weekly checks
of the nest boxes. Laying dates are calculated back
assuming that one egg per day is laid. All analyses of
laying dates and clutch sizes were restricted to first
clutches. Adults are caught during chick feeding and
identified by their alumi nium and colour-or ringed if not
previously caught. All nestlings are banded at an age of
7 days and weighed at 15 days, which is a good
approximation of their fledging weight. Laying dates are
given as April-days (1 April is April-day 1, 24 May is
April-day 54).
Temperature and other climatic data (precipitation,
wind speed, sunshine duration) for ‘De Bilt’, the main
weather station of the Royal Dutch Meteorological
Institute (KNMI), were used and obtained from the
KNMI’s website (http://www.knmi.nl/klimatologie/
daggegevens/download.cgi?language=eng).
Calibration of frass measurements
Caterpillar biomass was calculated from the raw frass
data using the formula of Tinbergen and Dietz (1994),
which essentially corrects the amount of frass produced
per unit biomass caterpillar for the temperature. To
calibrate these measurements we used data from
3 years (1994 1996) in which branches were sampled
from trees that also had caterpillar frass nets (n=60
tree-sample days). The main species of caterpillars
found were winter moth (Operophtera brumata ) and
oak leaf roller (Tortrix virirdana). When correlating the
biomass as recorded from the branch samples with the
biomass as calculated from the frass samples we used a
log transformation to normalise the data. We also used
a repeate d measures design with sample date repeated
over trees to take into account that the same trees were
sampled at different sampling dates (MIXED proce-
dure SAS).
The two measures of caterpillar biomass were highly
correlated (log biomass from branch samples versus log
biomass calculated from frass in a repeated measures
analysis of sample dates over trees: F
1,13
=19.35,
P<0.001). Biomass is expressed as gram caterpillar
biomass per m
2
per day, and thus ignores the height of
the tree under which the frass nets are placed. We
therefore include height in the subsequent analysis where
we correlate biomass as obtained from branch samples
with the biomass as calculated from the frass nets. There
were no additional significant explanatory variables,
neither of the tree characteristics (height and width at
breast height, F
1,50
<0.60, P>0.45) nor of sample date
characteristics (daily precipitation amount [Zandt 1994],
sunshine duration, daily mean wind speed and daily
mean temperature (note that the biomass as calculat ed
from frass was already corrected for temperature
according to Tinbergen & Dietz 1994; all F
1,3
<1.60;
P-values>0.30 in a step-wise analysis). We therefore,
used the biomass as calculated from the frass samples
(using the Tinbergen and Dietz 1994 formula) without
any further corrections as a measure for the biomass of
caterpillars in the trees.
Caterpillar biomass distribution
The distribution of the biomass of caterpillars over the
season cannot be fitted with a normal or any other dis-
tribution and hence we use three parameters to describe
the food peak: the peak date is the date at which the
largest biomass is recorded (also expressed in April-
days), the peak height is the largest biomass recorded in
a season (g biomass m
2
day
1
) and the peak width is
the number of days where the biomass is above 1 g
m
2
day
1
. The date at which the biomass rises above
and falls below this threshold value is estimated by fit-
ting a cubic poly nomial through the data points. In
3 years sampling either started after or ended before this
threshold was reached and hence for these years we
cannot estimate the peak width. Peak height and width
depend on the trees sampled and hence were only cal-
culated using the restricted set of seven trees that were
sampled every year (1993 2004). The biomass peak
date within an area is a much less tree-dependent mea-
sure and hence we used also the 1985 1992 period from
Verboven et al. (2001).
We constructed a descriptive model for the peak
biomass date (n=19 years, excluding 1991 as this was
a year with a late frost, which damaged the leaves of
the oak trees, Visser et al. 1998) by regressing the
peak dates against the mean temperature in all periods
of at least 10 day long between 1 January and the 31
May (in total 10,153 different periods) and selected the
period which had the highest correlation with peak
date. In the analysis of the spatial variation in cater-
pillar peak date, we used ANOVA and ANCOVA
models (GLM procedure SAS) with peak caterpillar
biomass date as dependent variable. Finally, we use
the descriptive model and the predicted temperatures
from an IPCC-SRES model, with an intermediate in-
crease in temperature (SRES-B2) (Esch 2005), and use
these to predict the caterpillar biomass distributions
for 2005–2100.
Great tit fitness parameters
Synchrony with food peak
The synchronisation between the birds’ breeding time
and the timing of the food peak was defined as the dif-
ference between the hatching date plus 9 days and peak
date (synchrony = hatching date + 9 peak date)
because great tit chicks grow fast at an age of 9 days and
food demands are then highest (Gebhardt-Henrich 1990;
Keller and van Noordwijk 1994).
166
Number and weight of fledglings
As part of our standa rd field protocol great tit chicks are
weighed to the nearest 0.1 g when they are 15 days old
and the number of fledged chicks is recorded by check-
ing the nest for dead chicks after fledging. For all broods
in the years 1985–2004 we calculated mean chick weight
per brood and the number of fledged chicks. For some
years clutch or brood size manipulations were carried
out. Since these experiments possibly affected chick
weight or the number of fledged chicks of a brood,
manipulated broods were excluded from this analysis.
The relationship between mean chick weight and the
number of fledged chicks with synchrony with the cat-
erpillar peak was analysed with GLMs in R 2.0.1. Full
models included the following explanatory variables and
all two-way interactions: synchrony, synchrony
2
, peak
height (maximum amount of caterpillar biomass re-
corded), peak width (number of days with caterpillar
biomass above 1 g m
2
day
1
) and brood size (number
of ha tched chicks). Identity link and normal error
structure was used for analysing mean chick weight and
log link and Poisson error structure for analysing the
number of fledged chicks. Minimum adequate models
were selected using step-wise backward deletion begin-
ning with the interactions.
Selection differentials
Selection differentials are defined as the covariance be-
tween trait and relative fitness and quantify the strength
of selection on the trait (e.g. Lande and Arnold 1983).
We used the number of offspring breeding in our pop-
ulation (termed recruits) as a fitness measure. Fitness
was converted to relative fitness by dividing the number
of recruits by the mean number of recruits in a given
year. To analyse whether the strength of selection is
related to the synchrony of the population with the
timing, the height or the width of the caterpillar peak,
we regressed the annual selection differentials against
these variables and all two-way interactions.
Predictive model of great tit laying dates
We used a Cox’s proportional hazards model (Cox 1972)
to build a descriptive model, based on the results of
Gienapp et al. (2005). This kind of mode l describes the
probability that an individual female will start with egg
laying as a function of an unspecified baseline hazard
and a set of explanatory variables. The explanatory
variables included in the model were female age (first
year breeder or older), temperature (calcula ted via a
‘linear predictor’), day length and the interaction be-
tween temperature and day length. To avoid very small
values of the baseline hazard and therefore possible
computational inaccuracies in the used algorithms, the
values for daily temperature and day length were
rescaled by subtracting their minimum value (i.e. effec-
tively rescaling the smallest value to 0).
Temperature and day length data
We used an IPCC-SRES model, with an intermediat e
increase in tempe rature (SRES-B2) (Esch 2005) as tem-
perature data for our prediction. The predictions of this
climate model are available at a temporal resolution of
1 day and a spatial resolution of 50*50 km
2
, in our case
around Arnhem (the Netherlands). Day length data for
every fifth year were obtained from the US Naval Office
via its website (http://www.aa.usno.navy.mil/data/docs/
RS_OneYear.html).
Simulation
The baseline hazard is only specified for days on which
a laying date was observed. To be also able to predict
laying dates outside this ‘time window’ we fitted an
exponential function to the baseline hazard estimated
by the proportional hazards mo del (r
2
=0.86,
P<0.0001) and used it for extrapolating the baseline
hazard. We calculated the hazard for all days from 1
January to 31 May for all years from 2005 to 2100
using the temperature and day length data, assuming
that 43% of the females are older than 1 year (which is
the average percentage for 1973–2004) and the
(extrapolated) baseline hazard. Using this ha zard we
simulated egg-laying dates of 100 females per year (see
Gienapp et al. 2005 for details). From the obtained
distribution of laying dates the annual mean simulated
laying dates were calculated.
Results
Describing the caterpillar biomass changes throughout a
season
In general, the peak in caterpillar biomass is narrow
given the length of the nestling period of about 17 days
as the number of days with a biomass above
1gm
2
day
1
is on average 24.3 days (n=9, range 19–
33.5). There is a correlation between the height of the
biomass peak and the timing: in late years the peak is
higher (Pearson’s r=0.76, P=0.004). There is no cor-
relation between the peak width and either the height or
the tim ing of the peak (both Pearson’s r<0.55;
P>0.13).
There is clear annual variation in the peak dates:
years differ when the caterpillar biomass peaks
(P<0.001, correcting for site, see below) and the range is
about 3 weeks, roughly equal to the wi dth of the peak.
There is also a significant advancement of the peak date
over the years 1985 2004 of 0.74 days a year
(F
1,17
=13.15, P=0.002 excluding 1991; F
1,18
=10.88,
P=0.004 incl. 1991; Fig. 1a, Visser et al, 1998, 2005).
167
There is no significant change in the width or the height
of the biomass peak over the years (both F
1,10
<0.75;
P>0.40).
There is clear spatial variation in the timing of the
biomass peak of oaks: some sites within the Hoge Vel-
uwe area are consistently early while others are late (site:
F
6,65
=17.52, P< 0.001; year: F
11,65
=22.35, P< 0.001;
Grieco et al. 2002). When site is replaced by character-
istics of the sampled trees (retaining year as a factor in
the model, F
11,34
=19.78, P<0.001), then height of the
tree and date of bud burst are not significant (both
F
1,32
<1.20; P>0.28), but there is a clear effect of the
width of the tree at breast heigth (F
1,34
=48.43,
P<0.001): for every 10 cm of the diameter of a tree the
biomass peak is 0.66 (SE 0.017) days earlier. Interest-
ingly, trees with a large diameter also have an early bud
burst (F
1,34
=61.54, P<0.001; again controlling for year
F
11,34
=4.91, P= 0.002): for every 10 cm of diameter the
bud burst date is 0.51 (SE 0.065) days earlier.
Constructing descriptive model of food phenology
Of all periods tested , the correlation between peak bio-
mass data and temperature was the highest for the per-
iod of 8th March to 17th of May (r
2
=0.78, P<0.001;
Fig. 1b; Visser et al. 1998, but note different tem perature
period). The width of the peak (i.e. the number of days
between caterpillar biomass above/below
1gm
2
day
1
) for which we have 9 years of data (1994,
1996, 1998–2004) also correlates with temperatu re. Here,
the period with the highest correlation is 16th of March
to 17th of May (r
2
=0.66, P=0.005).
There is also no interaction between site and tem-
perature (over the best fitting period; F
6,69
=0.42,
P=0.86), but sites differ in timing (F
6,75
=12.14,
P<0.001, with temperature in the analysis:
F
1,75
=141.44, P<0.001), the earliest and the latest site
differing by 9 days.
Correlation between great tit fitness parameters and
biomass distribution
Laying date
When we analyse the relationship between the timing of
the food peak and the start of egg laying by the birds
within the Hoge Veluwe area we find a correlation:
laying is earlier in years with an early food peak
(F
1,17
=4.52, P=0.048). Laying da te however only ad-
vances with 0.3 days for every day the food peak is
earlier (laying date = 8.23 + 0.30 · food peak date).
Verboven et al. (2001) have argued that the laying
dates of great tits in the Hoge Veluwe and the Oo-
sterhout areas differ in the way they correlate with the
caterpillar biomass peak dates but we do not find this
(effect of peak date on laying date: F
1,36
=16.91,
P<0.001 [combining the Oosterhout and Hoge Veluwe
data]), no effect of area · peak date interaction
(F
1,34
=0.81, P=0.37) nor of area (F
1,35
=0.01, P=0.98);
it should be noted that the food peak data for Oo-
sterhout came partly from much earlier years (1957–
1967) than for the Hoge Veluwe.
Synchrony with food peak
The phenology of both the birds, their laying date, and
of the food peak depend on temperature, but they differ
in the temperature periods they correlate best with.
However, if the temperatures used by the birds as cues to
time their reproduction, we would expect these temper-
atures to predict the time of optimal conditions to raise
their offspring, that is, caterpillar peak (Visser et al.
2004). We therefore used the temperatures for the best
fitting period for the laying dates (16 March 20 April,
r
2
=0.61 for 1973–2004) and determined how well these
temperatures predict the food peak. When we restricted
the analysis of the relationship between mean tempera-
ture and annual mean laying dates to the years for which
we have measured caterpillar peak dates, this relation-
ship still holds (r
2
=0.42, F
1,17
=12.22, P=0.003, no ef-
fect of year: F
1,16
= 0.26, P=0.62 or year ·
temperature interaction: F
1,15
=0.37, P=0.55). This
Fig. 1 Phenology of caterpillar biomass at the Hoge Veluwe (1985–
2004), in 1991 (open dot) a late frost damaged the fresh oak leaves
resulting in an extremely late peak date. a Advancement of
caterpillar peak date over time (broken line for all 20 years, solid
line with 1991 is excluded), and b caterpillar peak date (excluding
1991) versus the mean temperature from 8 March to 17 May (the
period that correlates best with the caterpillar phenology)
168
mean temperature over 16 March 20 April also cor-
relates with the timing of the food peak (r
2
=0.66,
F
1,16
=10.38, P=0.005) but there is an additional effect
of year (F
1,16
=11.18, P=0.004, no significant year ·
temperature interaction F
1,15
=1.26, P= 0.28). The sen-
sitivity of the timing of laying and of the food peak to
the temperatures over the period 16 March 20 April is
not significantly different (laying date: 3.34±0.97 day
C
1
, food peak: 4.01±1.25 day C
1
, F
1,32
=0.23,
P=0.63). This means that the temperatures used by the
birds as a cue are true predictors of the timing of the
food peak and it is therefore meaningful for the birds to
use them (Fig. 2). However, the year effect on the food
peak shows that the timing of the food peak is advanc-
ing: for the same mean temperature over the period 16
March 20 April the food peaks 0.57 days a year earlier.
We also have 19 years of data on the timing of the
food peak of Oosterhout, but 11 of these are from an
earlier period (1957–1967). We do find however also for
this area that the laying dates and the food peak corre-
late well with the temperatures from 16 March to 20
April (Fig. 2, temperature effect on laying date:
F
1,17
=107.30, P<0.001; temperature effect on food
peak: F
1,17
=16.17, P<0.001) and here also there is
no significant difference in the tempera ture sensitivity of
the two phenologial variables (laying date: 4.14±
0.40 day C
1
, food peak: 4.22±1.05 day C
1
,
F
1,32
=0.01, P=0.91). But in contrast to the Hoge Vel-
uwe, there is no significant year effect for the food peak
(F
1,16
= 0.76, P=0.40). This may reflect actual dif-
ferences between these areas but it is perhaps more likely
that this is due to the large number of years from the
earlier period for Oosterhout.
Clutch size
There was no effect of the height of the peak caterpillar
biomass on the mean annual clutch size (9.23 eggs;
F
1,10
=0.33, P=0.58) for the 12 years of the Hoge Vel-
uwe (1993–2004) where we sampled the same trees for a
number of years, making the peak height data compa-
rable over the years. The relationship between laying
date and clutch size varies significa ntly between years
(interaction between laying date and year:
F
18,2062
=3.99, P< 0.001 in a model with year and lay-
ing date, no significant quadratic terms) but the esti-
mates for the these annual slopes do not correlat e with
the biomass peak date (F
1,17
=3.46, P=0.08) as was
found for the pied flycatcher ( Ficedula hypoleuca ) at the
Hoge Veluwe area (Both and Visser 2005).
Reproductive success
The number of fledged chicks was strongly determined
by brood size (number hatched: V
2
1
=646.0, P<0.001;
n=1,368 broods) and by synchrony with the caterpillar
peak (Fig. 3a): when corrected for brood size, fewer
chicks fledged from broo ds raised before or after the
food peak (synchrony
2
: V
2
1
=97.0, P< 0.001; n=1,368
broods) and this relationship did not reach its maximum
at a synchrony of 0 days but 1.7 days later, that is, when
the chicks are 10.7 days old (synchrony: V
2
1
=10.9,
P<0.001).
Mean chick weight is significantly influenced by the
synchrony with the caterpillar peak (Fig. 3b): chicks
raised before or after the caterpillar peak are lighter
(Verboven et al. 2001). This effect is influenced by brood
size and the height of the caterpillar peak (synchrony
2
·
number hatched: F
1,529
=18.3, P<0.001; synchrony
2
·
peak height: F
1,529
=5.3, P=0.02): the effect of syn-
chrony on chick weight becomes weakened with an
increasing maximum caterpillar biomass but becomes
stronger with increasing brood size. The linea r syn-
chrony term of the model means that maximum chick
weight is not reached at 0 days synchrony (i.e. chicks are
9 days at the date of the biomass peak) but 1.1 days
later (when the chicks are 10.1 day old). This relation-
ship is also affected by the width of the caterpillar peak,
the wider is the peak the earlier becomes the date relative
to the caterpillar peak when chick weight is maximal
(synchrony · peak width: F
1,529
=7.1, P=0.007; cor-
recting for the brood size: F
1,529
=66.0, P<0.001).
Fig. 2 Laying date (open symbols) and peak biomass date (closed
symbols) for a Hoge Veluwe (1985–2004, excluding 1991) and b
Oosterhout (1957–1967, 1986–1988, 1996, 2001–2004) against the
mean temperature for 16 March 20 April (period that best
correlated with mean annual laying date). The dashed lines are the
regression lines for the laying dates. The solid lines are the
regression lines for the food peak as obtained from a model with
temperature and, for the Hoge Veluwe, year as explanatory
variables. The significant year effect for the Hoge Veluwe is
indicated by plotting the regression lines as obtained from the
model for 1985, 1995 and 2004. The difference in elevation for the
1985 and 2004 line is 10.9 days
169
Selection differentials
We could neither find a relationship between width nor
height of the caterpillar peak and annual selection dif-
ferentials (peak width: F
1,3
=0.02, P=0.90; peak height:
F
1,4
=0.0008, P=0.98). There was a trend of a negative
relationship between population mean synchrony and
the selection differ ential, however not significant
(F
1,15
=1.27, P=0.28).
Constructing descriptive model of laying date
Our proportional hazards model explained a significant
amount of variation in egg laying dates (Likelihood-ra-
tio = 1,549, df=4, P<0.001, see also Gienapp et al.
2005). All included explanatory vari ables were highly
significant (age: b=0 .35, P< 0.001; temperature:
b=1.60, P<0.001, day leng th: b=5.03, P<0.001, tem-
perature · day length: b=-1.03, P< 0.001). Older fe-
males laid earlier than young females (i.e. Perdeck and
Cave 1992; Wheelwright and Schultz 1994; Robertson
and Rendell 2001) as indicated by the positive regression
coefficient. Both temperature, calculated as ‘linear pre-
dictor’, and day length had a positive effect on the
probability that a female starts with egg laying (both
regression coefficients are positive). The negative
regression coefficient of the interaction between tem-
perature and day length means that the effect of tem-
perature decreases with increasing day length. Late in
the season, thus, lower temperatures trigger laying than
in early in the season (see also Gienapp et al. 2005).
Predicting future caterpillar biomass distributions
and laying dates
Over the period 2005–2100, caterpill ar peak dates are
predicted from the linear regression against spring
temperature and an IPCC-SRES temperature scenario.
Peak dates will advance by 0.20 days per year, which
will add up to a total advancement of 18 days (linear
regression: F
1,94
=50.0, P<0.001; Fig. 4). The predicted
width of the caterpillar peak will significantly decrease
(linear regression: b=0.13 ± 0.020, F
1,94
=40.0,
P<0.001).
From 2005 until 2100 mean laying dates are predicted
to advance by 0.16 days (±0.02) per year (linear
regression: F
1,94
=47.1, P<0.001) (Fig. 4). This amounts
to a total advance of 15 days over the whole period.
Simulated standard deviations show no significant time-
trend (linear regression versus year: F
1,94
=2.28,
P=0.13).
The predic ted time trends for laying dates in birds
and caterpillar phenology are however not significantly
different (linear regression: year · species interaction:
F
1,188
= 0.99, P=0.32).
Discussion
The optimal time for reproduction in the great tit is
clearly set by the time the biomas s of the caterpillars
peak: birds that have 11–12-day old chicks in their nest
at the annual peak date in biomass fledge the most
chicks (for their clutch size, Fig. 3a) and these are also
the heaviest (Fig. 3b), increasing the chance that they
will survive and breed (Verboven and Visse r 1998).
Fig. 3 Number of fledged chicks a and mean chick weight b of
great tits on the Hoge Veluwe in relation to synchrony with
caterpillar peak (synchrony=hatching date + 9 peak date). In a
fitted lines show relationship between number of fledged chicks and
synchrony for different brood sizes, in b the fitted line shows the
relationship between chick weight and synchrony for the average
number of hatched chicks, peak width and peak height
Fig. 4 Predicted phenology based on an IPCC-SRES scenario for
the Hoge Veluwe area. Annual means of predicted laying dates
(filled dots), and predicted dates when the caterpillar biomass
reaches its maximum (peak date) (open dots)
170
There is both spatial and annual variation in the time of
maximal caterpillar biomass and hence the optimal time
of breeding for the birds varies in space and time. The
annual variation in the date of the biomass peak is as
large as the width of the peak (about 3 weeks) em-
phasising the need for phenotypic plasticity.
One way the birds can cope with the spatial variation
is to learn when best to breed at the place they have
established themselves (Grieco et al. 2002). They cope
with the temporal variation by phenotypic plasticity of
their laying dates: the same individual lays at different
times in different years. Obviously, the birds need to
start reproduction quite some time before their chicks
are 11–12 days, that is, when they need to be syn-
chronised with the food peak. They, therefore, use cues
from their environment at the time of laying. We have
shown that laying dates correlate very well with tem-
perature (from 16 March to 20 April), and that the food
peak also correlates with these temperatures and that the
phenology of the birds and the caterpillar biomass re-
spond very similar to these temperatures. Hence, the
birds seem to be able to respond adequately to the
temporal variation in biomass peak date.
The observed mismatch between bird and caterpillar
biomass phenology at the Hoge Veluwe seems to be at
variance with the observation that both phenologies
respond in the same way to temperature. The reason for
this is that the ti ming of the food peak is also affected by
temperatures after 20 the April (note that the best fitting
period for the food peak is 8 March 17 May) and that
these temperatures have also increased (Visser et al.
1998). It turns out that this is more or less additive so
that the change of 1 C in the earlier period still leads to
a food peak that is 0.3 days earlier, but that the food
peak has advanced because the temperatures in the later
period have increased (see Fig. 2a). As the phenology of
the food is now earlier for the same temperatures over
the early period, but the phenology of the birds is not,
the consequence is that the interval between laying and
the food peak has become shorter. For a mean tem-
perature of 7 C over the period 16 Mar ch 20 April,
this interval was 37.5 days in 1985, but is only 26.6 days
in 2004 (a shift of 10.9 days). Given a clutch of ten eggs
(=10 days), 12 days of incubation and chicks of 12 days
when they should synchronise with the food peak
(=34 days), the result is that many of the birds have
their offspring in the nest too late to profit from the
short peak in caterpillar biomass. In the statistical model
this leads to a significant year (as a continuous variable)
effect in the analysis of the food peak.
The changed relation ship between early spring tem-
perature (16 March 20 April) and caterpillar peak date
will lead to selection for a changed reaction norm.
However, there is no need for the birds to become more
sensitive to temperature (the slope of the reaction norm)
but just that they have to lay earlier over the whole range
of temperatures (the inte rcept of the reaction norm). We
have indeed detected selection on the reaction norm for
the Hoge Veluwe popul ation, but surprisingly the
selection was stronger on the temperature sensitivity,
that is, the slope of the reaction norm, than on the laying
date in the average environment, that is, the intercept
(Nussey et al. 2005). At present, we cannot explain this
from the changes in the caterpillar biomass phenology as
this has not become more temperature sensitive (no
significant year · temperature interaction). We can only
propose several different explanations. A time series of
20 years may not be long enough to detect such an
interaction. For a meaningful description of reaction
norms it is crucial to identify the correct expla natory
variable, against which pheno types are regressed. Mean
temperatures are only a proxy for the real cues used by
the bir ds (Gienap p et al. 2005) and describing reaction
norms using a different explanatory variable may give
different results. Another explanation could be that
there are other selection pressures that select for a
steeper reaction norm, such as earlier settlements in
warmer years, and hence an earlier competition for
territories among the fledged offspring.
Great tits are facultative multi-brooders at the Hoge
Veluwe but over the past 20 years the proportion of
birds producing a second brood has strongly declined
(Visser et al. 2003). A potential explanation for this
decline would be a reduction in the width of the food
peak; a narrow food peak means a short time-window
for reproduction and hence fewer broods. However, we
cannot demonstrate a reduction in the width of the food
peak. Moreover, it is also likely that for the second
brood offspring caterpillars in oak are not the main food
source (Verboven et al. 2001).
Annual mean laying dates are correlated to annual
mean food peak dates, both on Oosterhout and the
Hoge Veluwe. This is what is expected if the phenotypic
plasticity in laying date is adaptive. Our results are in
contrast to those of Verboven et al. (2001) who claim
that there is such a relationship in Oosterhout and
Marley Wood (UK), where birds are generally not muli-
brooded, but not on the Hoge Veluwe and Vlieland
(NL), where part of the birds use to be double-brooded.
A major problem with the analysis of Verboven et al.
(2001) is, however, that they compare two areas for
which they have data from the fifties and sixties (Oo-
sterhout and Marley Wood) with two areas for which
they have data from the eighties and nineties (Hoge
Veluwe and Vlielan d), and find a difference. Given the
disrupted syn chrony in the recent two decades this could
well explain their results, rather than the incidence in
second broods in these two pairs of areas. In our anal-
ysis, with more recent years for both areas, there is no
longer a statistical differenc e between Oosterhout and
the Hoge Veluwe.
We have shown that the current reaction norm of
great tit laying date against temperature is no longer
adaptive. Given that laying dates reaction norms are
heritable (Nussey et al. 2005), we expect a response to
this selection and thereby a change in the reaction norm
of the birds. We predicted laying dates and food phe-
nology for 2005–2100 and found that both advance over
171
the next 100 years (Fig. 4) and are predicted to do so at
the same rate. However, this prediction is for a great tit
population in which no micro-evolution occurs. Obvi-
ously, what is needed is a predicted rate of change in
laying dates due to selection. Next, we should then
compare the rate of change in ecological conditions with
the rate of micro-evolution. In the absence of a change
in reaction norm we predict that the current mistiming
will not increase, which might mean that selection may
occur fast enough for the synchrony between the birds
and the caterpillar biomass phonologies to be restored,
and thereby reducing the negative impact of global cli-
mate change.
Acknowledgements We thank Jan Visser for maintaining the great
tit database, Ruben Smit for the measurements of the trees, Arie
van Noordwijk and many students for their help with the branch
sampling and Will Cresswell for his comments on the manu-
script. We are grateful to the board of the National Park de
Hoge Veluwe, to Barones van Boetzelaer van Oosterhout and the
State Forestry Service in Vlieland for the permission to work in
their woodlands.
References
Beebee TJC (1995) Amphibian breeding and climate. Nature
374:219–220
Both C, Visser ME (2005) The effect of climate change on the
correlation between avian life-history traits. Global Change
Biol 11:1606–1613
Brown JL, Li SH, Bhagabati N (1999) Long-term trend toward
earlier breeding in an American bird: A response to global
warming? Proc Natl Acad Sci USA 96:5565–5569
Crick HQP, Dudley C, Glue DE, Thomson DL (1997) UK birds
are laying eggs earlier. Nature 388:526–526
Daan S, Dijkstra C, Tinbergen JM (1990) Family planning in the
kestrel (Falco tinnunculus) the ultimate control of covariation
of laying date and clutch size. Behaviour 114:83–116
Dias PC, Blondel J (1996) Breeding time, food supply and fitness
components of Blue Tits Parus caeruleus in Mediterranean
habitats. Ibis 138:644–649
Durant JM, Anker-Nilssen T, Stenseth NC (2003) Trophic inter-
actions under climate fluctuations: the Atlantic puffin as an
example. Proc R Soc Lond Ser B Biol Sci 270:1461–1466
Esch M (2005) ECHAM4_OPYC_SRES_B2: 110 years coupled B2
run 6H values. DOI: 10.1594/WDCC/EH4_OPYC_SRES_B2
Fischbacher M, Naef-Daenzer B, Naef-Daenzer L (1998) Esti-
mating caterpillar density on trees by collection of frass drop-
pings. Ardea 86:121–129
Gebhardt-Henrich SG (1990) Temporal and spatial variation in
food availability and its effect on fledgling size in the great tit.
In: Blondel J, Gosler A, Lebreton J-D, McCleery R (eds)
Population biology of passerine birds, 2.0 edn. Springer, Berlin
Heidelberg New york, pp 175–186
Gienapp P, Hemerik L, Visser ME (2005) A new statistical tool to
predict phenology under climate change scenarios. Global
Change Biol 11:600–606
Grieco F, van Noordwijk AJ, Visser ME (2002) Evidence for the
effect of learning on timing of reproduction in blue tits. Science
296:136–138
Keller LF, van Noordwijk AJ (1994) Effects of local environmental
conditions on nestling growth in the great tit (Parus major L.).
Ardea 82:349–362
Klomp H (1970) Determination of clutch-size in birds a review.
Ardea 58:124
Lack D (1950) The breeding seasons of European birds. Ibis
92:288–316
Liebhold AM, Elkinton JS (1988a) Estimating the density of larval
gypsy-moth, Lymantria dispar (Lepidoptera, Lymantriidae),
populations using frass drop and frass production measure-
ments sources of variation and sample size. Environ Ento-
mol 17:385–390
Liebhold AM, Elkinton JS (1988b) Techniques for estimating the
density of late-instar gypsy-moth, Lymantria dispar (Lepidop-
tera, Lymantriidae), populations using frass drop and frass
production measurements. Environ Entomol 17:381–384
Martin TE (1987) Food as a limit on breeding birds a life-history
perspective. Ann Rev Ecol Syst 18:453–487
Naef-Daenzer L, Naef-Daenzer B, Nager RG (2000) Prey selection
and foraging performance of breeding Great Tits Parus major in
relation to food availability. J Avian Biol 31:206–214
Nussey DH, Postma E, Gienapp P, Visser ME (2005) Selection on
heritable phenotypic plasticity in a wild bird population. Sci-
ence 310:304–306
Parmesan C, et al (1999) Poleward shifts in geographical ranges of
butterfly species associated with regional warming. Nature
399:579–583
Parmesan C, Yohe G (2003) A globally coherent fingerprint of cli-
mate change impacts across natural systems. Nature 421:37–42
Pearce-Higgins JW, Yalden DW (2004) Habitat selection, diet,
arthropod availability and growth of a moorland wader: the
ecology of European Golden Plover Pluvialis apricaria chicks.
Ibis 146:335–346
Penuelas J, Filella I, Comas P (2002) Changed plant and animal life
cycles from 1952 to 2000 in the Mediterranean region. Global
Change Biol 8:531–544
Perrins CM (1970) The timing of bird’s breeding seasons. Ibis
112:242–255
Perrins CM (1991) Tits and their caterpillar food supply. Ibis
133(suppl):49–54
Thomas DW, Blondel J, Perret P, Lambrechts MM, Speakman JR
(2001) Energetic and fitness costs of mismatching resource
supply and demand in seasonally breeding birds. Science
291:2598–2600
Tinbergen JM, Dietz MW (1994) Parental energy-expenditure
during brood rearing in the great tit (Parus major) in relation to
body-mass, temperature, food availability and clutch size.
Funct Ecol 8:563–572
van Balen JH (1973) A comparative study of the breeding ecology of
the Great Tit (Parus major) in different habitats. Ardea 61:1–93
van Noordwijk AJ, McCleery R, Perrins C (1995) Selection for the
timing of great tit breeding in relation to caterpillar growth and
temperature. J Anim Ecol 64:451–458
Verboven N, Tinbergen JM, Verhulst S (2001) Food, reproductive
success and multiple breeding in the great tit Parus major. Ar-
dea 89:387–406
Verboven N, Visser ME (1998) Seasonal variation in local
recruitment of great tits: the importance of being early. Oikos
81:511–524
Verhulst S, van Balen JH, Tinbergen JM (1995) Seasonal decline in
reproductive success of the great tit variation in time or
quality. Ecology 76:2392–2403
Visser ME, et al (2003) Variable responses to large-scale climate
change in European Parus populations. Proc R Soc Lond Ser B
Biol Sci 270:367–372
Visser ME, Both C (2005) Shifts in phenology due to global climate
change: the need for a yardstick. Proc R Soc Lond Ser B Biol
Sci DOI:10.1098/rspb.2005.3356
Visser ME, Both C, Lambrechts MM (2004) Global climate change
leads to mistimed avian reproduction. Adv Ecol Res 35:89–110
Visser ME, Holleman LJM (2001) Warmer springs disrupt the
synchrony of oak and winter moth phenology. Proc R Soc
Lond Ser B Biol Sci 268:289–294
Visser ME, van Noordwijk AJ, Tinbergen JM, Lessells CM (1998)
Warmer springs lead to mistimed reproduction in great tits
(Parus major). Proc R Soc Lond Ser B Biol Sci 265:1867–1870
Zandt HS (1994) A comparison of 3 sampling techniques to esti-
mate the population-size of caterpillars in trees. Oecologia
97:399–406
172
... Seasonal windows of opportunity are intervals within a year that provide improved prospects for growth, survival, or reproduction . These seasonal windows reflect a favorable combination of biotic and abiotic factors in space and time, including periods of increased resource availability (e.g., Ogilvie et al., 2017;Visser et al., 2006), reduced predation pressure (e.g., Rasmussen & Rudolf, 2016;Urban, 2007), or more favorable climatic conditions (e.g., Bale et al., 2002;Hunter, 1993). Although the terminology has varied, seasonal windows of opportunity have long been recognized across a wide range of systems (Bale et al., 2002;Elton, 1927;Farzan & Yang, 2018;Hunter, 1993;Ogilvie et al., 2017;Rasmussen & Rudolf, 2016;Urban, 2007;Visser et al., 2006;Yang & Rudolf, 2010). ...
... These seasonal windows reflect a favorable combination of biotic and abiotic factors in space and time, including periods of increased resource availability (e.g., Ogilvie et al., 2017;Visser et al., 2006), reduced predation pressure (e.g., Rasmussen & Rudolf, 2016;Urban, 2007), or more favorable climatic conditions (e.g., Bale et al., 2002;Hunter, 1993). Although the terminology has varied, seasonal windows of opportunity have long been recognized across a wide range of systems (Bale et al., 2002;Elton, 1927;Farzan & Yang, 2018;Hunter, 1993;Ogilvie et al., 2017;Rasmussen & Rudolf, 2016;Urban, 2007;Visser et al., 2006;Yang & Rudolf, 2010). Conceptually, seasonal windows of opportunity represent a qualitative analog of the peaks in a continuous seasonal fitness landscape (Farzan & Yang, 2018;Yang & Rudolf, 2010), and recognize that seasonal periods of increased fitness commonly result from the combined effects of multiple bottom-up, top-down and abiotic factors that change over time (e.g., Farzan & Yang, 2018;. ...
Article
Full-text available
Seasonal windows of opportunity are intervals within a year that provide improved prospects for growth, survival, or reproduction. However, few studies have sufficient temporal resolution to examine how multiple factors combine to constrain the seasonal timing and extent of developmental opportunities. Here, we document seasonal changes in milkweed (Asclepias fascicularis)–monarch (Danaus plexippus) interactions with high resolution throughout the last three breeding seasons prior to a precipitous single‐year decline in the western monarch population. Our results show early‐ and late‐season windows of opportunity for monarch recruitment that were constrained by different combinations of factors. Early‐season windows of opportunity were characterized by high egg densities and low survival on a select subset of host plants, consistent with the hypothesis that early‐spring migrant female monarchs select earlier‐emerging plants to balance a seasonal trade‐off between increasing host plant quantity and decreasing host plant quality. Late‐season windows of opportunity were coincident with the initiation of host plant senescence, and caterpillar success was negatively correlated with heatwave exposure, consistent with the hypothesis that late‐season windows were constrained by plant defense traits and thermal stress. Throughout this study, climatic and microclimatic variations played a foundational role in the timing and success of monarch developmental windows by affecting bottom‐up, top‐down, and abiotic limitations. More exposed microclimates were associated with higher developmental success during cooler conditions, and more shaded microclimates were associated with higher developmental success during warmer conditions, suggesting that habitat heterogeneity could buffer the effects of climatic variation. Together, these findings show an important dimension of seasonal change in milkweed–monarch interactions and illustrate how different biotic and abiotic factors can limit the developmental success of monarchs across the breeding season. These results also suggest the potential for seasonal sequences of favorable or unfavorable conditions across the breeding range to strongly affect monarch population dynamics. Few studies have observed species interactions with sufficient temporal resolution to examine how multiple factors combine to constrain the seasonal timing and extent of developmental opportunities. Our results show early‐ and late‐season windows of opportunity for monarch recruitment and indicate that these windows were constrained by different combinations of factors. These findings show an important dimension of seasonal change in milkweed‐monarch interactions and suggest the potential for seasonal sequences of favorable or unfavorable conditions across the breeding range to strongly affect monarch population dynamics.
... In our worked example we speculated a 60 % decrease in breeding season but acclimation, changes in developmental time, phenological shifts or range shifts may occur in response to temperature increases (Visser et al., 2006;Carvalho et al., 2017;Davidson et al., 2021). While not able to predict the population response currently, the possible outcomes of range shifts or phenological shifts would probably result in trophic interaction changes which makes the RIP q method even more applicable for forecasting impact, should either occur. ...
Article
Full-text available
Predicting future changes in interspecific interactions continues to be a challenge for environmental managers. This uncertainty is exacerbated by increasing biological invasions and the likelihood that the strength of trophic interactions among native species will change. Abiotic variables influence predator resource utilisation and abundance as well as resource population dynamics. Currently no practical metric or impact prediction methodology can adequately account for all of these factors. Functional Response (FR) methods successfully incorporate resource utilisation rates with regards to resource density to quantify consumer-resource interactions under varying abiotic contexts. This approach has been extended to create the Relative Impact Potential (RIP) metric to compare invader vs native impact. However, this does not incorporate resource abundance dynamics, which clearly can also change with abiotic context. We propose a Resource Reproduction Qualifier (RRQ) be incorporated into the RIP metric, whereby RRQ is the reciprocal of the fraction or proportion to which reproduction (e.g. of prey species) changes under an environmental context. This modifies the RIP score to give a more informative RIP q value, which may be contextually increased or decreased. We empirically demonstrate the utility and benefits of including RRQ into impact potential predictions with an invasive species (the lionfish Pterois volitans) and two European native species (shanny fish Lipophyris pholis and lesser spotted dogfish Scy-liorhinus canicula) under different abiotic contexts. Despite high FR and abundance, lionfish impacts were reduced by increasing prey recruitment at higher temperatures, however, remained high impact overall. Shanny predatory impact increased with increasing temperature and was exacerbated by decreasing prey fecundity. Two population increase scenarios (50% and 80%) were assessed for lesser spotted dogfish under predicted temperature increases, preying upon E. marinus. Both scenarios indicated heightened predatory impact with increasing predator FR and decreasing prey fecundity. Our new metric demonstrates that accounting for resource reproductive responses to abiotic drivers, in tandem with the consumer per capita and abundance responses, better estimate the magnitudes of predicted inter-species interactions and ecological impacts. This can be used in stock assessments and predictions, as well as invasive species risk assessments in a comprehensive yet user-friendly manner..
... For instance, the NAO index may reflect insect abundance and phenology (Nott et al., 2002;Welti et al., 2020;Westgarth-Smith et al., 2012). The NAO can have considerable lagged effects on weather (Halkka et al., 2006), or there may be developmental time lags between weather conditions and the response in insect abundance (Visser et al., 2006). Thus, the effect of NAO during incubation may be acting on food availability during the important nestling growth stage. ...
Article
Full-text available
Environmental conditions during early-life development can have lasting effects shaping individual heterogeneity in fitness and fitness-related traits. The length of telomeres, the DNA sequences protecting chromosome ends, may be affected by early-life conditions, and telomere length (TL) has been associated with individual performance within some wild animal populations. Thus, knowledge of the mechanisms that generate variation in TL, and the relationship between TL and fitness, is important in understanding the role of telomeres in ecology and life-history evolution. Here, we investigate how environmental conditions and morphological traits are associated with early-life blood TL and if TL predicts natal dispersal probability or components of fitness in 2746 wild house sparrow (Passer domesticus) nestlings from two populations sampled across 20 years (1994-2013). We retrieved weather data and we monitored population fluctuations, individual survival, and reproductive output using field observations and genetic pedigrees. We found a negative effect of population density on TL, but only in one of the populations. There was a curvilinear association between TL and the maximum daily North Atlantic Oscillation index during incubation, suggesting that there are optimal weather conditions that result in the longest TL. Dispersers tended to have shorter telomeres than non-dispersers. TL did not predict survival, but we found a tendency for individuals with short telomeres to have higher annual reproductive success. Our study showed how early-life TL is shaped by effects of growth, weather conditions, and population density, supporting that environmental stressors negatively affect TL in wild populations. In addition, shorter telomeres may be associated with a faster pace-of-life, as individuals with higher dispersal rates and annual reproduction tended to have shorter early-life TL.
... Moreover, we conducted our study over a single breeding season, and certain effects of exposure to environmental factors or landscape features could change across years. For instance, the heat island effect may significantly increase across time within and between breeding seasons , which may affect prey phenology (e.g., Visser et al., 2006) and in return exacerbate its effect over breeding birds. Southern Barn Swallows are smaller and have longer breeding seasons than northern birds (Pagani-Núñez et al., 2016;Zhao et al., 2021), but since we were unable to take morphological measurements, or systematically record all breeding attempts in so many locations, we are unable to provide a more complete picture of the effect on fitness of exposure to urbanization and multiple environmental factors. ...
Article
Full-text available
In addition to landscape changes, urbanization also brings about changes in environmental factors that can affect wildlife. Despite the common referral in the published literature to multiple environmental factors such as light and noise pollution, there is a gap in knowledge about their combined impact. We developed a multidimensional environmental framework to assess the effect of urbanization and multiple environmental factors (light, noise, and temperature) on life-history traits and breeding success of Barn Swallows (Hirundo rustica) across rural to urban gradients in four locations spanning over 2500 km from North to South China. Over a single breeding season, we measured these environmental factors nearby nests and quantified landscape urbanization over a 1 km2 radius. We then analysed the relationships between these multiple environmental factors through a principal component analysis and conducted spatially explicit linear-mixed effects models to assess their effect on life-history traits and breeding success. We were particularly interested in understanding whether and how Barn Swallows were able to adapt to such environmental conditions associated with urbanization. The results show that there is significant variation in the exposure to environmental conditions experienced by Barn Swallows breeding across urbanization gradients in China. These changes and their effects are complex due to the behavioural responses ameliorating potential negative effects by selecting nesting sites that minimize exposure to environmental factors. However, significant relationships between landscape urbanization, exposure to environmental factors, and life-history traits such as laying date and clutch size were pervasive. Still, the impact on breeding success was, at least in our sample, negligible, suggesting that Barn Swallows are extremely adaptable to a wide range of environmental features.
... bird, climate change, phenological shifts, phenotypic plasticity, selection, timing of breeding after females invested in egg laying because the thermal environment during this time might directly affect chick survival and recruitment (Bonamour et al., 2019;Sauve et al., 2021;Visser et al., 2006). ...
Article
Many species have shifted their breeding phenology in response to climate change. Identifying the magnitude of phenological shifts and whether climate‐mediated selection drives these shifts is key for determining species’ resilience to climate change. Birds are a strong model for studying phenological shifts due to numerous long‐term research studies; however, generalities pertaining to drivers of phenological shifts will emerge only as we add study species that differ in life history and geography. We investigated 32 years of reproductive timing in a non‐migratory population of dark‐eyed juncos (Junco hyemalis). We predicted that plasticity in reproductive timing would allow females to breed earlier in warmer springs. We also predicted that selection would favour earlier breeding and asked whether the temperatures throughout the breeding season would predict the strength of selection. To test these predictions, we examined temporal changes in the annual median date for reproductive onset (i.e., first egg date) and we used a sliding window analysis to identify spring temperatures driving these patterns. Next, we explored plasticity in reproductive timing and asked whether selection favoured earlier breeding. Lastly, we used a sliding window analysis to identify the time during the breeding season that temperature was most associated with selection favouring earlier breeding. First egg dates occurred earlier over time and strongly covaried with April temperatures. Further, individual females that bred in more than one year, typically bred earlier in warmer Aprils, exhibiting plastic responses to April temperature. We also found significant overall selection favouring earlier breeding (i.e., higher relative fitness with earlier first egg dates) and variation in selection for earlier breeding over time. However, temperature across diverse climatic windows did not predict the strength of selection. Our findings provide further evidence for the role of phenotypic plasticity in shifting phenology in response to earlier springs. We also provide evidence for the role of selection favouring earlier breeding, regardless of temperature, thus setting the stage for adaptive changes in female breeding phenology. We suggest for multi‐brooded birds that advancing first egg dates likely increases the length of the breeding season, and therefore, reproductive success.
... For example, protandrous males may arrive at the breeding ground before a naturally selected optimum, with respect to temperature and resource availability (Brown & Brown, 2000;Irons et al., 2017). Due to climate change, early spring temperatures have increased (Høgda et al., 2013;Karlsen et al., 2007;The IPCC, 2013), leading to an earlier onset of the growing season (Høgda et al., 2013;Park et al., 2008) and earlier emergence of insects (Parmesan, 2006;Roy & Sparks, 2000;Visser et al., 2006). Thus, climate change has relaxed the constraints acting to oppose early spring arrival of migratory birds (Visser et al., 2015). ...
Article
Full-text available
Protandry is a widespread life-history phenomenon describing how males precede females at the site or state of reproduction. In migratory birds, protandry has an important influence on individual fitness, the migratory syndrome, and phenological response to climate change. Despite its significance, accurate analyses on the dynamics of protandry using data sets collected at the breeding site, are lacking. Basing our study on records collected during two time periods, 1979 to 1988 and 2006 to 2016, we aim to investigate protandry dynamics over 38 years in a breeding population of willow warblers (Phylloscopus trochilus). Change in the timing of arrival was analyzed in males and females, and protandry (number of days between male and female arrival) was investigated both at population level and within breeding pairs. Our results show advancement in the arrival time at the breeding site in both sexes, but male arrival has advanced to a greater extent, leading to an increase in protandry both at the population level and within breeding pairs. We did not observe any change in sex ratio that could explain the protandry increase, but pronounced temperature change has occurred and been reported in the breeding area and along the migratory route. Typically, natural selection opposes too early arrival in males, but given warmer springs, this counteracting force may be relaxing, enabling an increase in protandry. We discuss whether our results suggest that climate change has induced sex-specific effects, if these could be evolutionary and whether the timing of important life-history stages such as arrival at the breeding site may change at different rates in males and females following environmental shifts.
... Nestlings instead exclusively rely on caterpillars such that the great tit breeding phenology and success heavily depends on the availability and abundance of this food source (Visser et al. 2006). Great tits forage in the lower part of the canopy cover and on the ground where they can get infested by exophilic ticks such as Ixodes ricinus and Ixodes frontalis (Heylen et al. 2013a;Heylen and Matthysen 2010;Špitalská et al. 2011). ...
Article
What can developmental biology contribute toward mitigating the consequences of anthropogenic assaults on the environment and climate change? In this Spotlight article, we advocate a developmental biology that takes seriously Lynn Margulis' claim that ‘the environment is part of the body’. We believe this to be a pre-condition for developmental biology playing important roles in conservation and environmental restoration. We need to forge a developmental biology of the holobiont – the multi-genomic physiologically integrated organism that is also a functional biome. To this end, we highlight how developmental biology needs to explore more deeply the interactions between developing organisms, and their chemical, physical and biotic environments.
Article
Most of our understanding of the effects of climate warming on insect body size comes from laboratory experiments. Whether these studies predict patterns in nature is largely unknown. Here we examine the relevance of laboratory warming experiments for wild populations of the butterfly Pieris rapae. We tested two predictions: (i) butterflies reared at warmer temperatures in the laboratory should attain smaller adult sizes and have reduced flight ability, and (ii) in nature, this trait combination should lead to smaller butterflies visiting fewer flowers and accumulating less pollen. Overall, we found that warm‐reared butterflies were indeed smaller and flew more slowly compared to colder‐reared conspecifics. Additionally, wild‐caught small butterflies carried fewer, and a lower diversity of pollen grains compared to larger butterflies. Our warming experiments thus largely predicted pollen collection patterns in wild P. rapae. This study demonstrates that increased temperatures will likely have important consequences for butterfly‐plant interactions in nature. Here we used a laboratory experiment to demonstrate that warmer rearing temperatures reduced body size and fight speed in the widespread butterfly Pieris rapae. We also found that in nature, P. rapae of equivalent size to those reared in warmed laboratory conditions carried less pollen and visited fewer flower types compared to larger‐sized butterflies. Our study suggests that insect body size changes due to ongoing warming will have important effects on plant–pollinator interactions.
Article
Full-text available
Caterpillar frass dropping collections are compared with caterpillar density and biomass as estimated by branch samples taken from the same trees. A multiple regression model reveals (1) a linear relationship between mean caterpillar mass and frass dropping rate and (2) a non-linear effect of caterpillar density on frass dropping rates. At low density the probability that frass dropping from the trees is trapped in the collectors is reduced. The multiple approach explains 72% of the variance in the frass dropping rates. It therefore allows frass dropping rates to be convened into biomass estimates. Including air temperature and duration of sunshine per day does not significantly improve the model. Hence, we conclude that caterpillar size provides a good estimate to correct frass dropping rates for changing weather conditions. Analyses of within-tree and between trees variances reveal that frass dropping collection allows comparison of individual trees with a resolution of about 1 mg m-2 h-1. Since data on caterpillar growth (not abundance) can be obtained from a relatively small sample of branches, the conversion of frass dropping rates provides a relatively cheap and valuable tool to estimate caterpillar abundance in individual trees over large areas, even without using heavy equipment to reach the canopy.
Article
Full-text available
Naef-Daenzer, L., Naef-Daenzer, B. and Nager, R. G. 2000. Prey selection and foraging performance of breeding Great Tits Parus major in relation to food availability. - J. Avian Biol. 31: 206-214. We studied the nestling diet and the foraging performance of Great Tits in relation to prey abundance in the field. Numerous experimental studies present data on foraging decisions in captive Great Tits. Little is, however, known about prey selection in the field in relation to the food available and the consequences this has for the food delivery rate to nestlings. Since the foraging performance of the parents is one of the main determinants of fledging weight and juvenile survival, foraging behaviour is an important part of Great Tit reproduction. During the early breeding season up to 75% of the prey biomass delivered to the nestlings were spiders, which is in contrast with other studies. Only when caterpillars reached a size of 10-12 mg (approximately the average size of the spiders caught at that time) did the Great Tits change their preferences and 80-90% of the delivered prey masses were caterpillars, as reported by other authors. This 'switching' between prey occurred within a few days. It was not related to the changes in abundance but to size of caterpillars. The rate at which caterpillars were delivered to the nestlings (in mg:nestling:h) was strongly correlated with the caterpillar biomass available (in mg:m of branches) and nestling growth rate was significantly influenced by the mass of available caterpillars. The results provide evidence why perfect timing of breeding is so important for the Great Tit, and contribute to the understanding of the causal link between food supply, growth and breeding success.
Book
Population Biology of Passerine Birds is an up-to-date synthesis of the most recent developments in its field, especially in the framework of modern life history theories. Major topics include quantitative genetics, ecological physiology, the study of social structures using individually marked birds in the wild, and finally biometry, which allows to quantify such important parameters as survival at different stages of life and to create a model of the overall demography of the populations.
Chapter
Long-term studies on Great Tits in the Netherlands and England have revealed the importance of various environmental influences on the breeding biology of this species (see e.g. Perrins, 1965 and van Balen, 1973). Using Great Tits as a model I discuss the influences of temporal and spatial variation in food availability on fledgling size and weight. Temporal variation consisted of variation between three consecutive years as well as variation in hatching dates within years. Additionally, experimental brood size manipulations were performed to demonstrate the interaction between the main effects and the additional environmental conditions.
Article
We analysed the relationship between the timing of food availability and within-season variation of both reproductive success and nestling body size of Blue Tits Parus caeruleus in Mediterranean habitats. Synchronization between food supply and reproduction was expected to be positively related to fitness components. We measured deviation from maximum food supply using a parameter that we called “time-lag”, which quantifies the degree of synchronization between the date of maximum food requirements by the nestlings and the date of maximum caterpillar supply in the habitat. This parameter was expected to be related to reproductive success as measured by the number and body-condition of fledglings. The predictions were that time-lag should be negatively correlated with the proportion of nestlings raised to fledging and the size of the fledglings. These predictions have been tested in different types of habitat. The results demonstrate that caterpillar supply during a critical nestling period can have a strong influence on fitness components. As predicted, synchronization with caterpillar supply is positively related to the number and body size of fledglings. Since there is large between-habitat variation in the timing of food supply, the key issue seems to be the adjustment to local patterns of food availability.
Article
Climate change is apparent as an advancement of spring phenology. However, there is no a priori reason to expect that all components of food chains will shift their phenology at the same rate. This differential shift will lead to mistimed reproduction in many species, including seasonally breeding birds. We argue that climate change induced mistiming in avian reproduction occurs because there is a substantial period between the moment of decision making on when to reproduce and the moment at which selection operates on this decision. Climate change is therefore likely to differentially alter the environment of decision-making and the environment of selection. We discuss the potential consequences of such mistiming, and identify a number of ways in which either individual birds or bird populations potentially can adapt to reproductive mistiming.
Article
Daily weight increments of nestling Great Tits are expressed as ratios of observed increment divided by the increment expected under favourable conditions. We used this ratio to examine the effects of local environmental conditions on nestling growth. We demonstrate a positive relationship between nestling growth and food availability at that time and location. This relationship is stronger with the maximum rather than the median prey availability for the three to five trees sampled at a location. Residuals of the regression of the realized growth ratio on age are used to demonstrate a 10 to 20% reduction of growth on days with some daytime rain above one millimeter per hour, relative to growth in the same brood on dry days. Our trapping of the adults had an effect on nestling growth similar to that of two to three hours of rainfall. Finally, we show significantly later fledging in broods with slow growth. [KEYWORDS: Subsequent growth; breeding biology; sigmoid growth; body size; food; weight; survival; reproduction; swallow; model]