Effect of weather conditions on the spring migration of Eurasian Woodcock and consequences for breeding
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Ibis (2018) doi: 10.1111/ibi.12657
Effect of weather conditions on the spring migration
of Eurasian Woodcock and consequences for
breeding
KEVIN LE REST,1* ANDREW HOODLESS,2 CHRISTOPHER HEWARD,2 JEAN-LOUIS CAZENAVE3 &
1
YVES FERRAND
1
Office National de la Chasse et de la Faune Sauvage, 8 boulevard Albert Einstein, Nantes, 44300, France
2
Game & Wildlife Conservation Trust, Burgate Manor, Fordingbridge, Hampshire SP6 1EF, UK
3
Club National des Be cassiers, 105 rue Louis Pergaud, Villeneuve, Champniers, 16430, France
Migration is a critical period of time with fitness consequences for birds. The develop-
ment of tracking technologies now allows researchers to examine how different aspects
of bird migration affect population dynamics. Weather conditions experienced during
migration are expected to influence movements and, subsequently, the timing of arrival
and the energetic costs involved. We analysed satellite-tracking data from 68 Eurasian
Woodcock Scolopax rusticola fitted with Argos satellite tags in the British Isles and France
(2012–17). First, we evaluated the effect of weather conditions (temperature, humidity,
wind speed and direction, atmospheric stability and visibility) on migration movements
of individuals. Then we investigated the consequences for breeding success (age ratio)
and brood precocity (early-brood ratio) population-level indices while accounting for cli-
matic variables on the breeding grounds. Air temperature, wind and relative humidity
were the main variables related to migration movements, with high temperatures and
northward winds greatly increasing the probability of onward flights, whereas a trend
towards greater humidity over 4 days decreased the probability of movement. Breeding
success was mostly affected by climatic variables on the breeding grounds. The propor-
tion of juveniles in autumn was negatively correlated with temperature in May, but posi-
tively correlated with precipitation in June and July. Brood precocity was poorly
explained by the covariates used in this study. Our data for the Eurasian Woodcock indi-
cate that, although weather conditions during spring migration affect migration move-
ments, they do not have a major influence on subsequent breeding success.
Keywords: age ratio, Argos transmitter, breeding ground conditions, carry-over effects, early-
brood ratio.
The most common way for birds to cope with sea- wintering in habitat with higher resources, lower
sonal changes in weather and feeding conditions is predation risk and/or less competition (Alerstam
to make flights between separate wintering and et al. 2003, McKinnon et al. 2010). Undertaking
breeding grounds, a mechanism which has a strong migration, however, carries potential costs. It
genetic basis (Berthold et al. 1990, Berthold & encompasses an important part of the annual cycle
Helbig 1992, Pulido et al. 1996, Pulido 2007, (up to 40–60% for some long-distance migrants,
Liedvogel et al. 2011, Liedvogel & Lundberg Newton 2011) and comprises a period during
2014). The expression of this behaviour in a popu- which birds cross unknown areas with uncertain
lation is related to the fitness of migrating individ- food resources, weather conditions and predation
uals vs. resident ones, e.g. by breeding and/or risk. Conditions experienced during migration may
thus have major effects on the timing and
energetic costs of the journey (Jenni & Schaub
*Corresponding author.
Email: lerest.k@gmail.com 2003), as well as on individual survival (Klaassen
© 2018 British Ornithologists’ Union2 K. Le Rest et al. et al. 2014). For instance, food abundance and This study focuses on the effect of weather con- availability at stopover sites are known to be ditions on the pre-breeding migration of a cryptic important variables influencing the duration of bird, the Eurasian Woodcock Scolopax rusticola stopovers for refuelling (Newton 2006). Moreover, (hereafter Woodcock). It is a widespread species the weather conditions such as wind strength and living in woodland habitats. It has a broad diet, direction (Liechti 2006, Shamoun-Baranes et al. composed mainly of lumbricids, myriapoda and 2017) experienced during the trip strongly influ- arthropods. The spring migration of the Woodcock ence migration progress, depending on the flight alternates between flights of several hundred kilo- strategy of the species, i.e. soaring or flapping metres at night and stopovers for one or several flight, fuel deposit capacity and altitude of flight days (https://Woodcockwatch.com; http://www. (Alerstam 2011). becassesmigration.fr; Crespo et al. 2016). Food Pre-breeding migration has profound conse- abundance and availability play an important role quences for the breeding season. Indeed, it sets in habitat selection during winter (Duriez et al. important characteristics for breeding, such as the 2005) but are not expected to influence migration date of arrival at the breeding ground and body movements because most of the habitats crossed condition on arrival. The latter determines the during migration are suitable for Woodcock (Cre- capacity of the bird to invest energy directly in spo et al. 2016). However, weather conditions are breeding activity (Sandberg & Moore 1996) and expected to have a large influence on stopover the former constrains the timing of breeding, e.g. duration during Woodcock migration and, conse- breeding success may depend on the relationship quently, on the duration of the complete spring between hatching date and seasonal availability of migration. food for chicks (Both et al. 2005, Møller et al. To test this, we linked satellite tracking data 2008, Saino et al. 2011). The energetic cost of from 68 individuals with remote sensing datasets migration is especially important for species using available daily at a global scale. The effect of stored resources for breeding, i.e. capital breeders 14 climatic variables expected to affect bird (Stephens et al. 2009), but the timing of migration migration, related to wind, temperature, visibil- has consequences for breeding whatever the breed- ity, humidity and atmospheric conditions ing strategies of the species (Both & Visser 2001, (Richardson 1990a, Shamoun-Baranes et al. Newton 2006). 2017), and when relevant their trends over a The study of bird migration has been revolu- few days, were first tested to explain the migra- tionized by the miniaturization of geolocation tion movements of Woodcock at the individual technologies, which now enable the tracking of level. We then investigated the consequences of most birds throughout their travels. The large weather conditions experienced during migration amount of geolocation data recorded has revealed on the breeding success (proportion of juve- unexpected migration routes and behaviours (Bur- niles) and brood precocity (proportion of juve- ger & Shaffer 2008, Klaassen et al. 2014, niles having completed their post-juvenile Weimerskirch et al. 2015). This technology also moult) during the subsequent breeding season offers new opportunities to study the mechanisms at the population level. However, because the and factors driving migration strategy at the indi- energetic needs of breeding are thought to be vidual level. One of the main challenges lies in gained mainly from local resources, especially linking these geolocation data to relevant informa- for an income breeder such as Woodcock (see tion on the factors expected to influence bird Stephens et al. 2009), the timing and success of movements at relevant spatial and temporal scales breeding will also depend on climatic variables (e.g. meteorology, Shamoun-Baranes et al. 2010). on the breeding grounds (Guzman & Arroyo Remote sensing data appear to be the best solu- 2015). Temperature and precipitation on the tion, because it is almost impossible to realize breeding grounds were thus calculated and field surveys at each bird position (Gordo 2007). included in the analyses. We discuss the impli- Worldwide data are now easily available from cations of our findings for Woodcock population public datasets, e.g. land cover classification, vege- dynamics and, more generally, for our under- tation indices and climate datasets, which offer standing of carry-over effects of spring migration valuable information on parameters affecting bird conditions on the reproductive success of migra- migration. tory birds. © 2018 British Ornithologists’ Union
Weather conditions and Woodcock migration 3
Appendix S1 for details) and, when relevant, their
METHODS
trends over 4 days were extracted. The accuracy of
these variables was not available, which may have
Migration movements
introduced additional uncertainty into the analysis
From 2012 to 2017, 87 Woodcock were fitted (see Baker et al. 2017). All weather variables were
with Argos satellite tags (solar PPT Microwave available four times daily, but only information at
Telemetry, Inc., Columbia, MD, USA; 9.5 g) in 18 h (UTC) was extracted because Woodcock
February–March in Western Europe (63 in the Bri- migration movements started at dusk. Moreover, all
tish Isles in 2012–17 and 24 in France in 2015– altitude-dependent weather variables were
16). Tags were attached to the birds using a leg- extracted near the ground surface because GPS log-
loop harness (Rappole & Tipton 1991) composed gers fitted on Woodcocks showed that they mainly
of 1.6-mm-diameter, UV-resistant, marine-grade migrate at low altitude (A. Hoodless unpubl. data).
rubber cord (EPDM cord; Polymax Ltd, Bordon, Weather variables were averaged over 3 days (time
UK) passed through biomedical tubing (Silastic t, t + 1 and t + 2) to fit the tracking data time reso-
tubing; Cole Palmer, St Neots, UK). The tag lution (time t corresponded to the first 10-h ON
schedule alternated between a 10-h ON period period; time t + 1 and t + 2 to the 48-h OFF per-
(Argos messages transmitted every 60 s) and a 48- iod) and their trends were assessed over 4 days
h OFF period (no messages transmitted). Some of (adding the previous day, t 1).
the 87 tags failed prematurely and some others
failed to transmit any signal during spring migra-
Age and early-brood ratios
tion. Moreover, several Woodcock were resident in
the UK (no migration data) and some others died The effect of weather conditions experienced dur-
(hunted or unknown causes). Consequently, only ing spring migration on subsequent breeding ide-
68 tags from the 87 used provided meaningful ally would be investigated at the individual level.
data for studying spring migration. However, tracking data did not provide reliable
The spring tracking data were converted to a information about breeding success. Frequent
binary variable indicating whether the bird had movements of females during the breeding period
continued its migration (1) or not (0) between indicated incomplete incubation, but this was
two ON periods (first 10 h ON + 48 h OFF + insufficient to assess breeding success confidently.
next 10 h ON, i.e. 68 h elapsed time). A buffer of However, the proportion of juveniles (hereafter
50 km was defined to assess whether the distance age ratio) and the proportion of juveniles hav-
travelled between the two ON periods corre- ing finished their post-fledging moult before migra-
sponded to local movement or a migration flight. tion (hereafter early-brood ratio) were available
The threshold of 50 km was chosen according to each year from French and Russian monitoring
the home-range of Woodcock during the breeding schemes.
period, when movements were typically < 50 km Hunting of Woodcock wintering in Western
(mean sd = 14 23 km, n = 12). Such binary Europe is popular, especially on the wintering
data thus described the probability of migration grounds and during autumn migration. In France,
movement where at least two accurate locations a non-profit organization of Woodcock hunters
were transmitted during a 68-h period. In total, (Club National des Becassiers) collected wings of
729 spring locations from 68 individuals were used hunted Woodcock in order to determine the age
(Fig. 1). Spring migration movements in more and progress of moult (8000–10 000 wings col-
than 1 year were recorded for 31 individuals, with lected each year). At the same time, a ringing pro-
some birds yielding data for up to five consecutive gramme was managed by a governmental
years. organization (Office National de la Chasse et de la
Weather variables were extracted from the Faune Sauvage – ONCFS) and hunter organiza-
NCEP Reanalysis (Kalnay et al. 1996) 2.5° 9 2.5° tions (Federation Nationale des Chasseurs – FNC,
gridded data provided by the NOAA/OAR/ESRL Federations Departementales des Chasseurs – FDC).
PSD (Boulder, CO, USA) website (http://www. Thousands of birds (5000–7000) were ringed in
esrl.noaa.gov/psd/). Fourteen weather variables France each year through this programme and the
related to wind, temperature, humidity, visibility age and moult of individuals were assessed. Tens
and atmospheric conditions (see Table S1 in to a few hundred Woodcock were also caught in
© 2018 British Ornithologists’ Union4 K. Le Rest et al. Figure 1. The 729 locations from 68 individuals where spring migration movement of Woodcock could be assessed in a 68-h elapsed time. autumn near their breeding grounds in Central extracted from 1996 to 2017 for each location and Russia, thanks to a partnership between ONCFS date. As in the migration movement analysis, and a Russian organization (BirdsRussia). weather variables were averaged over 3 days (time Age (juvenile or adult) can be assessed easily t, t + 1 and t + 2) and their trends assessed over for Eurasian Woodcock using plumage characteris- 4 days (adding t 1). All values were then aver- tics (Clausager 1973, Ferrand & Gossmann 2009), aged yearly, which resulted in a unique set of and age ratio (proportion of juveniles) provided an weather conditions each year and enabled a correl- index of breeding success at the population level. ative analysis with the breeding indices. We con- The age ratio of birds caught in Central Russia in sidered three scenarios of change in migration early autumn was especially valuable because it phenology from 1996 to 2017: (1) no change; (2) gave a measure of breeding success at the start of a linear shift of ¼ day per year (based on mean autumn migration. However, the number of indi- spring passage at Helgoland; H€ uppop & H€ uppop viduals caught was low in comparison with the 2003); and (3) a non-linear shift driven by spring thousands of birds collected each year in France. precocity. Scenario (2) was evaluated by retrospec- The moult stage of juveniles in autumn/winter tively shifting migration dates by 1 day every gave important information on their hatching 4 years because weather conditions were only date because early-hatched juveniles (early extracted in the evening. Scenario (3) was broods) moulted all their secondary coverts, achieved using air temperature data extracted for whereas late-hatched juveniles (late broods) scenario (1) to give an index of rise in spring tem- showed some juvenile coverts (autumn migration perature at the flyway scale. An annual delay or stops moult of the wing feathers, Ferrand & Goss- advancement in spring temperatures was then cal- mann 2009). The proportion of early-brood juve- culated. This variable was scaled, as the overall niles therefore provided information on breeding trend from 1996 to 2017 corresponded to a gen- phenology and its possible consequences for eral shift of ¼ day per year, as in scenario (2). The breeding success, e.g. higher mortality rate in difference in temperature values was then con- chicks reared later. verted into a number of days, which allowed a To link such population-level indices with non-linear shift of dates to be considered. spring migration weather, the tracking data loca- Age and early-brood ratios were also expected tions and dates were used as a sample to represent to be influenced by weather on the breeding Woodcock flyway movements. Important weather grounds. If so, failing to account for these vari- variables found to affect migration movement were ables may lead to fallacious results due to © 2018 British Ornithologists’ Union
Weather conditions and Woodcock migration 5
spurious correlations between weather during during migration on age and early-brood ratios,
spring migration and the breeding period. were investigated using linear models with normal
Monthly mean temperature and precipitation in errors. Model residuals were visually checked for
May, June and July from 1996 to 2017 were normality. A linear model accounted for overdis-
therefore extracted from known breeding loca- persion due to the independent variance parame-
tions of 63 tracked Woodcocks. Total precipita- ter r of the error term, whereas the use of a
tion from January to April was extracted as an binomial model would imply that variance was a
index of pre-breeding soil–water conditions each function of the mean, which ignores overdisper-
year. sion. An alternative would be to use quasi-bino-
mial models, but AIC could not be strictly
evaluated with such an approach. Note that the
Statistical analyses
results remained equivalent using either quasi-
The effect of weather conditions on spring binomial or linear models. The source of data
migration movements was examined using gener- (birds ringed in Russia, ringed in France or hunted
alized linear mixed models, with a binomial in France) was used as a factor to account for the
error term and a logit link function. Weather differences between these sources, ensuring that
variables were added as fixed effects and the covariates explained relative changes. Results
individual bird was specified as a random effect. remained unchanged whether data source was
Candidate covariates (fixed effects) were chosen considered as a random or a fixed effect. Parame-
according to prior knowledge or their expected ters were estimated by weighted least squares to
effect on migration movements. It was not possi- account for the differences in the number of indi-
ble to consider all subsets of these candidate viduals ni used to calculate age and early-brood
covariates (2k possibilities with linear effects). ratios.
Covariate selection was thus done in several Candidate climatic variables were those
steps starting from a model including only tem- selected in the migration movement analysis
perature and wind effects (strong support from (eight variables from the best model selected)
prior knowledge). In total, about 150 subsets of and conditions at the breeding grounds (seven
covariates were considered (weather conditions variables, see above). These covariates could not
and individual-based variables, e.g. age and all be considered in the same model because the
breeding location) while avoiding collinearity number of covariates considered was high rela-
between covariates (cross correlation < 0.7, see tive to the number of observations, and because
Dormann et al. 2013). All subsets of covariates of collinearity between some of them. The analy-
were then classified according to Aka€ıke’s infor- sis was thus carried out considering two main
mation criterion (AIC) and we retained the best potential drivers of variation in observed index
model and alternative models having DAIC < 2, values: (1) the climatic variables extracted on
i.e. those having ‘substantial support’ according the breeding grounds and (2) the weather condi-
to Burnham and Anderson (2002). Row coeffi- tions experienced during migration. The second
cients of the models and their uncertainty are potential driver of variation was investigated by
provided in Appendix S2 (see Table S2) and testing the three scenarios detailed above on the
averaged parameters were calculated from this changes in migration phenology since 1996, i.e.
subset of models (Symonds & Moussali 2011). (scenario 1) no change in migration dates,
Model selection was performed on a subsample (scenario 2) a shift of ¼ day per year and
of 668 observations for which there were no (scenario 3) an annual change in migration dates
missing values for any of the covariates. How- depending on air temperature. Covariate selection
ever, coefficients were estimated from the maxi- was done independently for each of these four
mum number of observations with no missing sets of covariates (see Tables S3 and S7, in
value in the set of covariates considered (from Appendices S3 and S4, respectively). Model selec-
668 to 729 depending on the covariates consid- tion was carried out using a corrected AIC (AICc)
ered). Model residuals were checked for spatial with n = 23 (number of studied years). AICc was
correlation. compared between the different scenarios, which
The effects of climatic variables extracted on allowed assessment of the main cause of variation
the breeding grounds, and the weather conditions in age ratio and early-brood ratio.
© 2018 British Ornithologists’ Union6 K. Le Rest et al.
wetness. Precipitation rate had a skewed distribu-
RESULTS
tion, which prevented a good evaluation of its
effect on migration movements.
Migration movements
Migration movements more frequently occurred
Variable selection based on AIC resulted in reten- in the presence of positive air surface temperatures
tion of 15 models with DAIC < 2. The models (> 3 °C), stable atmospheric conditions (positive
had from 10 to 15 parameters (Table 1, best four-layer lifted index) and favourable winds
Appendix S2). Air surface temperature, winds (especially northward winds, Fig. 2). A quadratic
(especially northward wind), atmospheric stability effect was fitted to atmospheric conditions, such
and relative humidity trend were the main that the highest values resulted in a decrease in
weather variables associated with migration move- movement probability, but the number of observa-
ments (Table 1, Fig. 2). These variables were pre- tions with such extreme conditions was small and,
sent in all of the 15 competing models, and their hence, the uncertainty was high. Conversely,
coefficients were very similar across them (see migration movements occurred less frequently in
Appendix S2). Air pressure and relative humidity the presence of an increasing trend in relative
were selected in most of the alternative models, humidity. Movements were most frequent in the
but the estimated coefficients showed high uncer- presence of a slightly decreasing or stable relative
tainties. Neither cloud cover nor cloud ceiling humidity trend. Uncertainty was very high for
height (visibility) were selected in the models. The conditions with a strongly decreasing relative
same was true for vertical wind speed and soil humidity trend (Fig. 2). Bird breeding origin (lon-
gitude and/or latitude of breeding area, linear
migration distance; Table 1) also showed weak
Table 1. Averaged parameters from the best set of models positive correlations with probability of migration
(15 models with DAIC < 2, see Table S2 in Appendix S2) on movement. Longitude of the breeding area (Fig. 2)
the probability of migration movement.
was the variable which best highlighted this corre-
lation, with reduced uncertainty in most of the
Averaged
Covariates Averaged standard models considered (Table 1, Appendix S2).
(scaled) coefficients errors t-values Selected interactions (lifted index 9 air pressure
trend and eastward 9 northward wind) enhanced
Intercept 0.60 0.15 3.95
AIC values, but coefficient estimates showed high
Longitude of 0.23 0.11 2.03
breeding area uncertainty.
Air temperature 0.74 0.13 5.73
Eastward wind 0.15 0.09 1.66
(u-component) Age and early-brood ratios
Northward wind 0.50 0.10 4.90
The proportion of young among late autumn
(v-component)
Air pressure 0.15 0.10 1.46 and winter birds collected in France was lower
Air pressure trend 0.14 0.10 1.40 than that of Russian birds caught in early
Best four-layer 0.23 0.12 1.90 autumn. Moreover, age ratio differed between
lifted index ringed and hunted individuals in France, but
(Best four-layer 0.25 0.07 3.55
both showed synchrony over time (Fig. 3a,
lifted index)²
Relative humidity 0.15 0.12 1.27 r = 0.79, n = 21 years). All the models consid-
Relative humidity 0.14 0.09 1.53 ered accounted for these differences through
trend inclusion of the factor ‘type of data’. Variable
(Relative humidity 0.16 0.06 2.71 selection showed that the lowest AIC was
trend)²
achieved by a model accounting for the climatic
(Eastward wind 0.13 0.09 1.55
9 Northward wind) variables during the breeding period (Table 2,
Best four-layer lifted 0.14 0.10 1.44 Table S3 in Appendix S3). The alternative best
index 9 Air pressure model, accounting only for weather conditions
trend during spring migration, was that of the non-
Latitude of breeding 0.12 0.11 1.03
shifted migration hypothesis, but it performed
area
Migration distance 0.17 0.12 1.42 very poorly (DAIC > 30, Table S4). The most
important variables correlating with age ratio
© 2018 British Ornithologists’ UnionWeather conditions and Woodcock migration 7
Figure 2. Effect of air temperature (°C), atmospheric stability (best four-layer lifted index), wind speed (m/s), relative humidity trend
and longitude of breeding area (decimal degree) on migration movements. The grey line with shaded area represent the mean and
its 95% confidence interval. The dashed horizontal line corresponds to a probability of migration movement of 0.5.
were temperature in May and precipitation in positively related and stabilized in the presence
July. The former was negatively related to the of high precipitation (Table 2, Fig. 4). Age ratio
proportion of juveniles, whereas the latter was was also higher in the presence of high
© 2018 British Ornithologists’ Union8 K. Le Rest et al. Figure 3. Trends in (a) age ratio and (b) early-brood ratio from 1996 to 2017. Each source of data is drawn using a different colour (see key). temperature and precipitation in June. Overall, n = 21 years). Values in the Russian data were climatic variables on the breeding grounds slightly higher (Fig. 3b) but between-year variation explained 65% of the observed variation in age was high, so selection based on AIC favoured ratio (calculated from the deviance ratio of the models without the factor ‘type of data’ model including only the factor ‘source of (Table S4_1 in Appendix S4). The lowest AIC in data’). In contrast, the models accounting for the models examining early-brood ratio was weather conditions during migration only achieved by a model accounting for weather con- explained 10% (non-linear shifted migration), ditions under the hypothesis of non-shifted migra- 15% (linear shifted migration) and 37% (non- tion. Air pressure and relative humidity showed shifted migration) of the variation in age ratio negative correlations with early-brood ratio (Tables S3_1 to S3_3 in Appendix S3). More (Table 3), but their effects on migration move- importantly, the main weather variables corre- ments had high uncertainty (Table 1, lated with age ratio were not the ones that Appendix S2). The models evaluated under the were strongly related to migration movements. alternative hypotheses of linear and non-linear In particular, coefficients estimated from air shifted migration phenology performed poorly pressure and air pressure trend exhibited high (Appendix S4). Of the climatic variables extracted uncertainties (Table 1). at the breeding grounds, only temperature in May Early-brood ratio was quite similar between was selected. Early-brood ratio was higher in the hunted and ringed individuals in France (r = 0.48, presence of higher temperatures in May. © 2018 British Ornithologists’ Union
Weather conditions and Woodcock migration 9
Table 2. Estimated parameters of the best model fitting age correlated with spring migration movements of
ratio to the climatic variables at the breeding grounds (see Woodcock (Fig. 2), with individuals logically tak-
Appendix S3 for other models, all having DAIC ≫ 2).
ing advantage of tailwinds. Intriguingly, the effect
of northward wind was much stronger than the
Estimated
Estimated standard effect of eastward wind, although there was no
Covariates (scaled) coefficients errors t-values apparent advantage to Woodcock in heading north
faster. Wind speed and direction may not only
Intercept 59.27 0.83 71.41
facilitate migration flights but may also provide
Source: ringed 16.94 3.40 4.98
in Russia some information on conditions at the following
Source: hunted 7.35 0.73 10.07 stopover sites. In particular, northward winds may
in France give important information about temperature
Temperature May 2.50 0.41 6.10 trend (and snowmelt) at a broad scale, whereas
Precipitation July 2.89 0.40 7.22
eastward winds are associated with rainy weather
Precipitation June 1.69 0.41 4.12
Temperature June 0.75 0.45 1.67 and atmospheric instability (which may generate
(Temperature June)² 1.57 0.37 4.24 snow when reaching cold areas). Another explana-
(Precipitation July)² 1.54 0.46 3.35 tion might be related to the structure of winds in
R²: 0.80 AICc: 397.2 the different layers of the atmosphere. Indeed,
birds may adapt their flight altitude by a compro-
mise between wind support and air temperature
(Kemp et al. 2013). We need reliable data on the
DISCUSSION
flight altitude of tracked Woodcock to investigate
this hypothesis.
Migration movements
Atmospheric conditions have not often been
Spring migration movements of Eurasian Wood- shown to influence non-soaring bird migration
cock were clearly associated with particular (Richardson 1990b). Yet, in theory, stable atmo-
weather conditions. Air surface temperature spheric conditions could also be advantageous for
showed the strongest positive correlation, but flapping-flight migration by reducing turbulence in
wind, atmospheric stability and relative humidity the air and the risk of encountering storms, thus
trend were also highly related to migration move- encouraging the bird to migrate. Atmospheric sta-
ment probability (Fig. 2). Many studies of bird bility was evaluated here by the best four-layer
migration have shown that temperature has a lifted-index and was the third most important vari-
major role in migration schedules (e.g. Marra et al. able (after temperature and wind) explaining
2005, H€ uppop & Winkel 2006, Gordo 2007, Woodcock spring migration movements (based on
Tøttrup et al. 2010). Woodcock in particular estimated coefficients, Table 1). Moreover, as
avoided moving when temperatures were below or expected, Woodcock migration movements were
near 0 °C, and migration movement frequency lower when relative humidity increased. We were
reached 50% above 4 °C and 80% above 11 °C. unable to find any study in which this variable has
Temperature may have an indirect effect on previously been related to migration movements.
migration movements, i.e. increasing food abun- Both atmospheric stability and relative humidity
dance and availability at new sites along the migra- trend should thus be given greater consideration in
tion route, which consequently stimulates birds to future studies modelling bird migration movements.
continue their travel. This would be analogous to Migration movements of Woodcock still
the green wave hypothesis for goose migration (Si occurred when weather conditions were not opti-
et al. 2015, and references therein). mal. Indeed, the movement probability was rarely
Wind is also known to be an important factor lower than 0.2 for each single effect (Fig. 2). Thus,
influencing migration movements (Liechti 2006, it seems that only a combination of unfavourable
Sinelschikova et al. 2007). Biologging technologies elements could drastically reduce migration move-
permitting the tracking of birds throughout their ment probability towards zero and, conversely, just
migration travels have recently highlighted the a few favourable factors could encourage some
importance of this variable at the individual level birds to continue spring migration. A large amount
(Conklin & Battley 2011, Safi et al. 2013). In our (88%) of the observed variance (evaluated from
study, wind direction and strength were strongly the null model with only the individuals
© 2018 British Ornithologists’ Union10 K. Le Rest et al.
Figure 4. Effect of temperature (°C) in May and June and precipitation (mm) in June and July on age ratio. The grey line with
shaded area represent the mean and its 95% confidence interval.
Table 3. Estimated parameters of the best model fitting early- accuracy according to the density of weather sta-
brood ratio to the weather conditions under the hypothesis of tions, which may create additional variance or
non-shifted migration phenology (see Appendix S4 for other even bias in analysis (Baker et al. 2017). Another
models, all having DAIC > 2). part of this unexplained variance results from vari-
ables and processes unaccounted for in the models.
Estimated
For instance, energy expenditure and body condi-
Covariates Estimated standard
(scaled) coefficients errors t values tion could affect the decision of Woodcock to stop
or fly on at the individual level throughout the
Intercept 63.89 0.47 135.94 migration. Variables related to visibility (cloud
Air pressure 1.98 0.45 4.4
cover and cloud-ceiling height) were not shown to
Relative humidity 1.56 0.47 3.32
(Air pressure)² 0.97 0.30 3.23 affect Woodcock migration movements.
R²: 0.33 AICc: 379.01 We found no difference in migration move-
ments between juvenile (second calendar year)
and adult Woodcock (third or more calendar
considered as a random effect) remained unex- year). However, Woodcock breeding further east
plained by the best model. A part of this unex- had a higher rate of migration movements than
plained variance may be related to the accuracy of birds breeding closer to the wintering grounds
the variable used in the model. In particular, some (Table 1). This is mainly related to the migration
variables may have spatial variation in their distance the birds have to cover, as the longitude
© 2018 British Ornithologists’ UnionWeather conditions and Woodcock migration 11
of the breeding area is highly correlated with the thus increase post-fledging mortality. Guzman and
distance between the wintering and breeding areas Arroyo (2015) also showed a negative effect of tem-
(r = 0.97, n = 82 spring travels, DAIC of the alter- perature in July and a positive effect of precipitation
native model = 1.1, see Appendix S2). Woodcock on relative Woodcock abundance in Spain during
breeding further east and undertaking longer the following winter, which would be partly related
migrations thus compensate for their long-distance to the breeding success that summer. However,
travel by increasing their rate of migration move- they also found a positive effect of precipitation in
ment. Migration will have a greater impact on the May, but did not provide interpretation of this
fitness of these individuals, and they may need to result.
be more efficient at finding resources at stopover We evaluated breeding success using the age
areas and/or their breeding grounds might be ratio, which only gave the proportion of young
expected to be of higher quality. Woodcock in the population a few months after
the breeding period. The limitations in the inter-
pretation of results on age ratio must be borne in
Age and early-brood ratios
mind because, although breeding productivity will
Age ratio was higher among hunted Woodcock have a strong influence on age ratio values, condi-
than those captured for ringing, but both trends tions during autumn migration may affect adults
across years were very similar. The higher age ratio and juveniles differently, e.g. differential mortality
among hunted birds is likely to be due to differen- and/or migration routes. The abundance of Wood-
tial distribution of young birds according to hunt- cock in Western Europe in a given winter is likely
ing pressure, such that young birds reaching to depend strongly on weather conditions (espe-
wintering areas for the first time are more likely to cially frost) in northern and eastern countries,
settle in hunted areas by accident than are adults, which push Woodcock towards southern areas
which almost always returned to the same winter- such as Spain (Peron et al. 2011). Unfortunately,
ing location (Fadat 1991). Once accounting for the this was not taken into account by Guzman and
factor ‘source of data’, annual age ratio was mostly Arroyo (2015), and hence the relative importance
related to climatic variables on the breeding of breeding ground conditions and autumn/winter
grounds. Conversely, weather conditions during weather (and their correlations) on the abundance
spring migration were unlikely to influence the pro- of wintering Woodcock still need to be evaluated.
portion of juveniles. Temperature in May and June In contrast to age ratio, the early-brood ratio was
and precipitation in June and July showed strong poorly related to the covariates studied. The weather
correlations with age ratio. In May, most female variables during spring migration that were
Woodcock on the Russian breeding grounds are correlated with early-brood ratio were not those cor-
incubating, so a higher temperature might be con- related with migration movements. Such inconsisten-
strued as good for breeding. However, temperature cies were probably due to multicollinearity between
and precipitation were positively correlated in May the covariates measured during spring migration and
(r = 0.43, n = 23 years) and heavy precipitation other important variables unaccounted for in our
might reduce hatching success. The negative corre- models. Of the variables extracted on the breeding
lation between age ratio and temperature in May grounds, only temperature in May may have a weak
suggests that unstable weather in May (high tem- positive influence on early-brood ratio. This is plausi-
peratures and precipitation) was detrimental to ble, as high temperatures in May advance spring phe-
hatching success. Conversely, in June, most young nology and hence hatching date (Hoodless &
would have hatched and would benefit from high Coulson 1998). However, because temperature in
temperatures and precipitation, which was consis- May was negatively correlated with age ratio, it
tent with the correlations we found. In July, dry might produce antagonistic effects on breeding suc-
weather will probably drive earthworms deeper into cess and breeding precocity.
the soil and make it harder for Woodcock to probe,
which could prevent them from feeding on soil
Implications for population dynamics
invertebrates (Peach et al. 2004, Hoodless & Hirons
2007). Low rainfall in July could particularly affect In a recent review, Shamoun-Baranes et al. (2017)
Woodcock chicks, most of which will be full grown provided an overview of the effect of atmospheric
by July but still relatively na€ıve foragers, and could conditions on bird migration at the individual level
© 2018 British Ornithologists’ Union12 K. Le Rest et al.
and gave some insights into the consequences of by the Shooting Times Woodcock Club. ONCFS and
these conditions at the population level. The CNB financed the tags in France. CNB financed the
authors highlighted that the feedback of weather tags thanks to a membership fee as well as private
donations. We would like to thank Bruno Meunier
conditions experienced during migration on popu-
(President of CNB) and all the members of CNB for
lation dynamics has not yet been sufficiently inves- their input since the first stages of this project. Owen
tigated. We have shown that the weather Williams (The Woodcock Network) and Luke Harman
conditions experienced during spring migration (University College Cork) helped to fit tags in Wales
had little influence on breeding success (age ratio and Ireland. Pierre Launay (CNB), Damien Coreau and
after the breeding period) or breeding phenology Francßois Gossmann (ONCFS), as well as many ringers
(early-brood ratio) of Woodcock. Breeding success from Reseau Becasse (ONCFS/FNC/FDC) helped to
catch Woodcock and fit tags in France. One anony-
was mainly correlated with climatic variables on
mous reviewer, the associate editor, James Pearce-Hig-
the breeding grounds, which could influence both gins, and the editor, Dan Chamberlain, provided
hatching success and the survival of juveniles. Pop- valuable comments and corrections which greatly
ulation dynamics of Eurasian Woodcock wintering improved the manuscript.
in Western Europe are thus more likely to be
related to the conditions on the breeding grounds
than to conditions experienced during the spring
REFERENCES
migration, which is consistent with the income
breeding strategy of this species (Stephens et al. Alerstam, T. 2011. Optimal bird migration revisited. J.
2009). Although many studies have examined Ornithol. 152(S1): 5–23.
Alerstam, T., Hedenstrom, A. & Akesson, S. 2003. Long-
carry-over or delayed effects of previous conditions distance migration: evolution and determinants. Oikos 103:
on breeding success, most have also concluded that 247–260.
conditions on the breeding grounds were the most Baker, D.J., Hartley, A.J., Pearce-Higgins, J.W., Jones,
important (Legagneux et al. 2012, Ockendon et al. R.G. & Willis, S.G. 2017. Neglected issues in using weather
2013, but see Finch et al. 2014). and climate information in ecology and biogeography.
Divers. Distrib. 23: 329–340.
Although spring migration weather conditions Bauthian, I., Gossmann, F., Ferrand, Y. & Julliard, R. 2007.
were not shown to affect breeding success of Eura- Quantifying the origin of Woodcock wintering in France. J.
sian Woodcock wintering in Western Europe at the Wildl. Manage. 71: 701–705.
population level, they could affect the fitness of Bell, C.P. 2000. Process in the evolution of bird migration and
birds at the individual or sub-population level. pattern in avian ecogeography. J. Avian. Biol. 31: 258–265.
Berthold, P. & Helbig, A.J. 1992. The genetics of bird
Indeed, indices used here reflected what was hap- migration: stimulus, timing, and direction. Ibis 134(S1): 35–40.
pening on average, but may have hidden differences Berthold, P., Wiltschko, W., Miltenberger, H. & Querner, U.
between individuals or between geographical 1990. Genetic transmission of migratory behavior into a
regions. Breeding success should ideally be evalu- nonmigratory bird population. Experientia 46: 107–108.
ated at the individual level to assess further its link Both, C. & Visser, M.E. 2001. Adjustment to climate change
is constrained by arrival date in a long distance migrant bird.
with spring migration conditions. In particular, birds Nature 411: 296–298.
breeding further from their wintering areas (east of Both, C., Bijlsma, R.G. & Visser, M.E. 2005. Climatic effects
the Ural Mountains) would theoretically be the on timing of spring migration and breeding in a long-
most impacted by weather conditions during migra- distance migrant, the Pied Flycatcher Ficedula hypoleuca. J.
tion. These birds represented a small proportion of Avian. Biol. 36: 368–373.
Burger, A.E. & Shaffer, S.A. 2008. Application of tracking
the population studied here, probably because of and data-logging technology in research and conservation of
evolutionary processes exerting selection pressure seabirds. Auk 125: 253–264.
on migration (Bell 2000, Liedvogel et al. 2011, Sha- Burnham, K.P. & Anderson, D.R. 2002. Model Selection and
moun-Baranes et al. 2017). However, the propor- Multimodel Inference: A Practical Information-theoretical
tion of these long-distance migrants is higher than Approach, 2nd edn. New York: Springer.
Clausager, I. 1973. Age and sex determination of the
previously thought (see Bauthian et al. 2007). It Woodcock, Scolopax rusticola. Danish Rev. Game Biol. 8:
would be interesting to evaluate whether their 1–18.
occurrence in winter in Western Europe has chan- Conklin, J.R. & Battley, P.F. 2011. Impacts of wind on
ged over recent decades and, if so, why. individual migration schedules of New Zealand Bar-tailed
Godwits. Behav. Ecol. 22: 854–861.
Crespo, A., Rodrigues, M., Telletxea, I., Iba n~ ez, R., Dıez, F.,
GWCT funded the study in Britain and Ireland, thanks
Tobar, J.F. & Arizaga, J. 2016. No habitat selection during
mainly to private donations for tags and funds raised
© 2018 British Ornithologists’ UnionWeather conditions and Woodcock migration 13
spring migration at a meso-scale range across mosaic the flight altitude of nocturnal migrants in mid-latitudes. Ibis
landscapes: a case study with the Woodcock (Scolopax 155: 734–749.
rusticola). PLoS One 11: e0149790. Klaassen, R.H.G., Hake, M., Strandberg, R., Koks, B.,
Dormann, C.F., Elith, J., Bacher, S., Buchmann, C., Carl, Trierweiler, C., Exo, K.M., Bairlein, F. & Alerstam, T.
G., Carre , G., Garcıa Marque z, J.R., Gruber, B., 2014. When and where does mortality occur in migratory
Lafourcade, B., Leita ~o, P.J., Mu
€ nkemu€ ller, T., McClean, birds? Direct evidence from long-term satellite tracking of
C., Osborne, P.E., Reineking, B., Schro € der, B., Skidmore, raptors. J. Anim. Ecol. 83: 176–184.
A.K., Zurell, D. & Launtenbach, S. 2013. Collinearity: a Legagneux, P., Fast, P.L.F., Gauthier, G. & Bety, J. 2012.
review of methods to deal with it and a simulation study Manipulating individual state during migration provides
evaluating their performance. Ecography 36: 27–46. evidence for carry-over effects modulated by environmental
Duriez, O., Ferrand, Y., Binet, F., Corda, E., Gossmann, F. conditions. Proc. R. Soc. B Biol. Sci. 279: 876–883.
& Fritz, H. 2005. Habitat selection of the Eurasian Liechti, F. 2006. Birds: Blowin’ by the wind? J. Ornithol. 147:
Woodcock in winter in relation to earthworms availability. 202–211.
Biol. Cons. 122: 479–490. Liedvogel, M. & Lundberg, M. 2014. The genetics of
Fadat, C. 1991. Age-ratio des tableaux de chasse de migration. In Hansson, L.-A. & Akesson, S. (eds) Animal
be casses (Scolopax rusticola). Signification biologique et Movement Across Scales: 219–231. New York: Oxford
utilisation pour la bonne gestion des populations University Press.
be cassie res. Bulletin Mensuel de l’ONC – Nume ro Spe
cial Liedvogel, M., Akesson, S. & Bensch, S. 2011. The genetics
Scientifique et Technique, 141–172. of migration on the move. Trends Ecol. Evol. 26: 561–569.
Ferrand, Y. & Gossmann, F. 2009. Ageing and sexing series Marra, P.P., Francis, C.M., Mulvihill, R.S. & Moore, F.R.
5: ageing and sexing the Eurasian Woodcock Scolopax 2005. The influence of climate on the timing and rate of
rusticola. Wader Study Group Bull. 116: 75–79. spring bird migration. Oecologia 142: 307–315.
Finch, T., Pearce-Higgins, J.W., Leech, D.I. & Evans, K.L. McKinnon, L., Smith, P.A., Nol, E., Martin, J.L., Doyle, F.I.,
2014. Carry-over effects from passage regions are more Abraham, K.F., Gilchrist, H.G., Morrison, R.I.G. & Be ^ ty, J.
important than breeding climate in determining the breeding 2010. Lower predation risk for migratory birds at high
phenology and performance of three avian migrants of latitudes. Science 327: 326–327.
conservation concern. Biodivers. Conserv. 23: 2427–2444. Møller, A.P., Rubolini, D. & Lehikoinen, E. 2008.
Gordo, O. 2007. Why are bird migration dates shifting? A Populations of migratory bird species that did not show a
review of weather and climate effects on avian migratory phenological response to climate change are declining.
phenology. Clim. Res. 35: 37–58. Proc. Natl Acad. Sci. USA 105: 16195–16200.
Guzma n, J.L. & Arroyo, B. 2015. Predicting winter Newton, I. 2006. Can conditions experienced during migration
abundance of Woodcock Scolopax rusticola using weather limit the population levels of birds? J. Ornithol. 147: 146–166.
data: implications for hunting management. Eur. J. Wildl. Newton, I. 2011. Migration within the annual cycle: species,
Res. 61: 467–474. sex and age differences. J. Ornithol. 152: 169–185.
Hoodless, A.N. & Coulson, J.C. 1998. Breeding biology of Ockendon, N., Leech, D. & Pearce-Higgins, J.W. 2013.
the Woodcock Scolopax rusticola, in Britain. Bird Study 45: Climatic effects on breeding grounds are more important
195–204. drivers of breeding phenology in migrant birds than
Hoodless, A.N. & Hirons, G.J.M. 2007. Habitat selection and carryover effects from wintering grounds. Biol. Let. 9:
foraging behaviour of breeding Eurasian Woodcock 20130669.
Scolopax rusticola: a comparison between contrasting Peach, W.J., Denny, M., Cotton, P.A., Hill, I.F., Gruar, D.,
landscapes. Ibis 149: 234–249. Barrit, D., Impey, A. & Mallard, J. 2004. Habitat selection
Hu€ ppop, O. & Hu € ppop, K. 2003. North Atlantic Oscillation by Song Thrushes in stable and declining farmland
and timing of spring migration in birds. Proc. R. Soc. Lond. populations. J. Appl. Ecol. 41: 275–293.
B Biol. Sci. 270: 233–240. ron, G., Ferrand, Y., Gossmann, F., Bastat, C.,
Pe
Hu€ ppop, O. & Winkel, W. 2006. Climate change and timing of Guenezan, M. & Gimenez, O. 2011. Nonparametric spatial
spring migration in the long-distance migrant Ficedula regression of survival probability: visualization of population
hypoleuca in central Europe: the role of spatially different sinks in Eurasian Woodcock. Ecology 92: 1672–1679.
temperature changes along migration routes. J. Ornithol. Pulido, F. 2007. Phenotypic changes in spring arrival:
147: 326–343. evolution, phenotypic plasticity, effects of weather and
Jenni, L. & Schaub, M. 2003. Behavioural and physiological condition. Clim. Res. 35: 5–23.
reactions to environmental variation in bird migration: a Pulido, F., Berthold, P. & van Noordwijk, A.J. 1996.
review. In Berthold, P., Gwinner, E. & Sonnenschein, E. Frequency of migrants and migratory activity are genetically
(eds) Avian Migration: 155–171. Berlin: Springer. correlated in a bird population: evolutionary implications.
Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, Proc. Natl Acad. Sci. USA 93: 14642–14647.
D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, Rappole, J.H. & Tipton, A.R. 1991. New harness design for
J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., attachment of radio transmitters to small passerines. J. Field
Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J., Orn. 62: 335–337.
Leetmaa, A., Reynolds, B., Jenne, R. & Joseph, D. 1996. Richardson, W.J. 1990a. Timing of bird migration in relation
The NCEP/NCAR 40-year reanalysis project. Bull. Am. to weather: updated review. In Gwinner, E. (ed.) Bird
Meteor. Soc. 77: 437–470. Migration: 78–101. Berlin: Springer.
Kemp, M.U., Shamoun-Baranes, J., Dokter, A.M., van Richardson, W.J. 1990b. Wind and orientation of migrating
Loon, E. & Bouten, W. 2013. The influence of weather on birds: a review. Experientia 46: 416–425.
© 2018 British Ornithologists’ Union14 K. Le Rest et al.
Safi, K., Kranstauber, B., Weinzierl, R., Griffin, L., Rees, in behavioural ecology using Akaike’s information criterion.
E.C., Cabot, D., Cruz, S., Proan ~ o, C., Takekawa, J.Y., Behav. Ecol. Sociobiol. 65: 13–21.
Newman, S.H., Waldenstro € m, J., Bengtsson, D., Kays, R., Tøttrup, A.P., Rainio, K., Coppack, T., Lehikoinen, E.,
Wikelski, M. & Bohrer, G. 2013. Flying with the wind: scale Rahbek, C. & Thorup, K. 2010. Local temperature fine-
dependency of speed and direction measurements in tunes the timing of spring migration in birds. Integr. Comp.
modelling wind support in avian flight. Move. Ecol. 2013: 1–4. Biol. 50: 293–304.
Saino, N., Ambrosini, R., Rubolini, D., von Hardenberg, J., Weimerskirch, H., Delord, K., Guitteaud, A., Phillips, R.A. &
Provenzale, A., Hu € ppop, K., Hu€ ppop, O., Lehikoinen, A., Pinet, P. 2015. Extreme variation in migration strategies
Lehikoinen, E., Rainio, K., Romano, M. & Sokolov, L. between and within wandering albatross populations during
2011. Climate warming, ecological mismatch at arrival and their sabbatical year, and their fitness consequences. Sci.
population decline in migratory birds. Proc. R. Soc. Lond. B Rep. 5: 8853.
Biol. Sci. 278: 835–842.
Sandberg, R. & Moore, F.R. 1996. Fat stores and arrival on Received 11 September 2017;
the breeding grounds: reproductive consequences for revision accepted 7 August 2018.
passerine migrants. Oikos 77: 577–581. Associate Editor: James Pearce-Higgins.
Shamoun-Baranes, J., Bouten, W. & van Loon, E. 2010.
Integrating meteorology into research on migration. Integr.
Comp. Biol. 50: 280–292. SUPPORTING INFORMATION
Shamoun-Baranes, J., Liechti, F. & Vansteelant, W.M.G. 2017.
Atmospheric conditions create freeways, detours and tailbacks Additional supporting information may be found
for migrating birds. J. Comp. Physiol. A. 203: 509–529. online in the Supporting Information section at
Si, Y., Xin, Q., de Boer, W.F., Gong, P., Ydenberg, R.C. & the end of the article.
Prins, H.H.T. 2015. Do Arctic breeding geese track or overtake
a green wave during spring migration? Sci. Rep. 5: 8749.
Appendix S1. Details on the weather variables
Sinelschikova, A., Kosarev, V., Panov, I. & Baushev, A. used to study migration movements.
2007. The influence of wind conditions in Europe on the Appendix S2. Covariate selection, parameter
advance in timing of the spring migration of the Song estimates and model averaging for migration move-
Thrush (Turdus philomelos) in the south-east Baltic region. ment probability.
Int. J. Biometeorol. 51: 431–440.
Stephens, P.A., Boyd, I.L., McNamara, J.M. & Houston, A.I.
Appendix S3. Covariate selection and para-
2009. Capital breeding and income breeding: their meaning, meter estimates for age ratio.
measurement, and worth. Ecology 90: 2057–2067. Appendix S4. Covariate selection and para-
Symonds, M.R.E. & Moussali, A. 2011. A brief guide to meter estimates for early-brood ratio.
model selection, multimodel inference and model averaging
© 2018 British Ornithologists’ UnionYou can also read
