Content uploaded by Yannis Vardanis
Author content
All content in this area was uploaded by Yannis Vardanis on Dec 16, 2019
Content may be subject to copyright.
Consistency in long-distance bird migration: contrasting patterns in
time and space for two raptors
Yannis Vardanis
a
,
*
, Jan-Åke Nilsson
a
, Raymond H. G. Klaassen
b
,
c
,
d
,
e
, Roine Strandberg
a
,
Thomas Alerstam
a
a
Evolutionary Ecology, Department of Biology, Lund University, Sweden
b
Dutch Montagu's Harrier Foundation, Scheemda, The Netherlands
c
Animal Ecology Group, Centre for Evolutionary and Ecological Studies (CEES), University of Groningen, The Netherlands
d
Dutch Centre for Avian Migration and Demography, Netherlands Institute of Ecology (NIOO-KNAW), Wageningen, The Netherlands
e
Department of Animal Ecology, Netherlands Institute of Ecology (NIOO-KNAW), Wageningen, The Netherlands
article info
Article history:
Received 8 May 2015
Initial acceptance 1 July 2015
Final acceptance 2 November 2015
Available online
MS. number: 15-00393R
Keywords:
bird migration
consistency
individual variation
marsh harrier
osprey
satellite telemetry
As the evolutionary responses to environmental change depend on selection acting on individual dif-
ferences, disentangling within- and between-individual variation becomes imperative. In animal
migration research, multiyear tracks are thus needed to estimate the individual consistency of pheno-
typic traits. Avian telemetry studies have recently provided the first evidence of individuality across
space and time in animal migration. Here, we compare repeatability patterns of routes and timing be-
tween two migratory birds, the marsh harrier, Circus aeruginosus, and the osprey, Pandion haliaetus,as
recorded by satellite tracking. We found interspecific contrasts with low repeatability in timing and
duration and a high repeatability in routes for ospreys, but the reverse pattern for marsh harriers. This
was mainly caused by (1) larger between-individual variation in routes for ospreys (broad-front
migration) than for marsh harriers (corridor migration) and a higher degree of repeated use of the same
stopover sites among ospreys, and (2) higher within-individual consistency of timing and duration
among marsh harriers, while individual ospreys were more flexible. Our findings suggest that in-
dividuality in space and time is not a shared trait complex among migrants, but may show adaptive
variation depending on the species' life history and ecology.
©2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
The potential among migratory animals to respond to environ-
mental change is a topic that so far has mostly been studied at the
population level for logistical reasons (Knudsen et al., 2011;
Sutherland, 1998). However, as the evolutionary responses of a
population depend on selection pressures acting on the phenotypic
and genetic variation among individuals (Lynch &Walsh, 1998),
disentangling individual variation will be necessary for predicting
how a changing environment might affect migration patterns
(Conklin, Battley, &Potter, 2013). To this end, multiyear tracking of
individual migratory histories are needed (Alerstam, 2006; Baker,
1978; Pulido, 2007a), allowing us to quantify the consistency of
phenotypes.
The individual consistency of animal behaviours (Bell, Hankison,
&Laskowski, 2009) has typically been estimated by the
repeatability index r, the part of phenotypic variation in a popula-
tion that can be attributed to differences between individuals
(Nakagawa &Schielzeth, 2010). The degree of route fidelity of long-
distance migration has mainly been assessed for avian migrants,
although estimates exist also for marine turtles (Broderick, Coyne,
Fuller, Glen, &Godley, 2007; Schofield et al., 2010). Nevertheless,
the study of consistency in migratory behaviour is still critically
defined by the difficulty of tracking individual migratory organisms
over long periods. Thus, most studies have focused on calculating
repeatability for time-related traits in one or a few annual stages
(Table 4 in Thorup, Vardanis, Tøttrup, Kristensen, &Alerstam,
2013).
Recently it has become possible to evaluate the repeatability in
both space and time for long-distance migration across the entire
annual cycle by tracking individual birds during several years using
satellite telemetry (spanning 2e7 years: Lopez-Lopez, Garcia-
Ripolles, &Urios, 2014; Vardanis, Klaassen, Strandberg, &Aler-
stam, 2011) and light level-based geolocation (spanning 2e3 years:
Dias, Granadeiro, &Catry, 2013; Stanley, MacPherson, Fraser,
*Correspondence: Y. Vardanis, Department of Biology, Lund University, Ecology
Building, SE-223 62 Lund, Sweden.
E-mail address: yannis.vardanis@gmail.com (Y. Vardanis).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
http://dx.doi.org/10.1016/j.anbehav.2015.12.014
0003-3472/©2015 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 113 (2016) 177e187
McKinnon, &Stutchbury, 2012). In this context, comparing re-
peatabilities among species can be particularly informative, as the
repeatability for a trait may vary among species, and these differ-
ences can provide insights into the ecological conditions shaping
consistency in migratory traits (van Noordwijk et al., 2006).
Here, we use long-term tracking data to compare the spatial and
temporal patterns of repeatability in migratory routes and timing
between two long-distance migrants, the marsh harrier, Circus
aeruginosus, and the osprey, Pandion haliaetus. Both species migrate
between Sweden and sub-Saharan West Africa within the same
flyway, in which adults are highly faithful to individual breeding
and wintering sites (Cramp &Simmons, 2004). They both migrate
using a combination of flapping and soaring flight and have similar
resulting total migration speeds and moult strategies (Alerstam,
Hake, &Kjell
en, 2006; Strandberg et al., 2008). At the same time,
they show ecological differences in their migratory strategies, such
that ospreys migrate on a broad front (Fransson et al., 2001;Poole,
1989), consistently revisit individual-specific intermediate goal
areas (Alerstam et al., 2006) and travel during a shorter time
window of the day (Mellone et al., 2012). Furthermore, marsh
harriers have a more complex migration schedule as they tend to
make premigratory stopovers (Strandberg et al., 2008).
Our null hypothesis was that the two species would show the
same pattern of high repeatability in migratory timing and low
repeatability in migratory route found in some recent studies of
long-distance terrestrial bird migrants (e.g. Lopez-Lopez et al.,
2014; Stanley et al., 2012; Vardanis et al., 2011). Studies of repeat-
ability in migratory animals are still very few, making it important
to consider the alternative hypothesis that repeatability patterns
may differ between species, which would suggest that individual
flexibility in migratory timing and routes are traits that show
adaptive variation depending on the species' life history and ecol-
ogy. Thus, our aim was to test this hypothesis by comparing the
degree of individual consistency in migratory timing and routes
between two raptor species with similar breeding and wintering
areas. We consider possible ecological reasons related to stopover
behaviour and broad- versus narrow-front migration for these
interspecific differences in the individuality of bird migration.
METHODS
Tracking Data
We used tracking data from eight adult ospreys and six adult
marsh harriers that made more than one migratory round trip (an
autumn and a subsequent spring journey) between the breeding
sites in Sweden and the wintering sites in West Africa and back to
Sweden again during the years 1996e2012, as recorded by satellite
telemetry. Part of this data set has been published (ospreys:
Alerstam et al., 2006; marsh harriers: Vardanis et al., 2011), but we
add more data for both species in this comparative study. The
complete data set consisted of eight ospreys and 38 tracks (three
individuals and 21 tracks added), with 21 autumn and 17 spring
journeys, as well as six marsh harriers and 39 tracks (eight tracks
added), with 21 autumn and 18 spring journeys (Table 1,Fig. 1). We
used several types of satellite transmitters. The very first devices
deployed in 1996 on ospreys lasted for a limited time period
(normally only one annual cycle) and provided locations only every
3 or 6 days (Argos PTT-100, Microwave Telemetry Inc., Columbia,
MD, U.S.A.). In 1997e2006, we used solar-powered transmitters to
track ospreys and marsh harriers (Solar Argos PTT-100, Microwave
Telemetry Inc.), which provided locations on a (near-) daily basis
(for details, see: Hake, Kjell
en, &Alerstam, 2001; Strandberg et al.,
2008). From 2006 we also used GPS-based satellite transmitters
(Solar Argos/GPS PTT-100, Microwave Telemetry Inc.), which
provide bihourly GPS positions during the day. In four cases, the
transmitter failed and the bird was fitted with a new transmitter
(second tracking period; Table 1). All transmitters were tracked by
the ARGOS system (CLS, Toulouse, France). For battery- and solar-
powered transmitters, geographical positional fixes were calcu-
lated using Doppler shift, and individual positions differ in accuracy
(location accuracy classes Z, B, A, 0e3; see www.argos-system.org/
manual). We inspected the entire data set to exclude nonvalidated
(Z) and low-accuracy locations (0, A, B) obviously off-track (single
positions implying a combination of travel speed of >25 m/s and
off-course direction deviating >45
off the route inferred by the
high-quality positions before and after them). Positions were
transformed to the Mercator map projection (Gudmundsson &
Alerstam, 1998). Following the same approach as Vardanis et al.
(2011), we used these coordinates to calculate the longitudes and
dates that the birds' migratory routes crossed the latitudes 46
N,
36
N and 26
N. These three latitudes were selected to represent
measures of timing and geography of the migratory routes for the
crossing of different regions: Europe, the Mediterranean Sea and
the Sahara Desert, respectively.
Analyses
We calculated the repeatability of the longitudes and dates us-
ing the ANOVA-based method (Lessells &Boag, 1987; Nakagawa &
Schielzeth, 2010) for a maximum of eight ospreys and six marsh
harriers for variables with at least two measurements (Table 1).
This analysis served to test the degree of individual consistency in
longitudes, timing and duration across years, for the two species.
The duration of the migratory journey was calculated as the total
time for the journey between breeding and winter sites excluding
pre- and postmigratory movements within 2
latitude zones close
to the breeding and wintering areas (Strandberg et al., 2008). We
corrected durations for the small variation in latitudinal extent of
the journeys (individuals had their breeding sites as well as winter
sites at slightly different latitudes) by dividing by the number of
latitude degrees crossed during the journey, and using this measure
Table 1
Tracking records obtained by satellite telemetry during 1996e2012 of repeated
migratory journeys made by eight adult ospreys and six adult marsh harriers be-
tween breeding sites in Sweden and wintering sites in West Africa
Species Individual Number of complete tracks Tracking period
Autumn Spring 1st 2nd
Start End Start End
Osprey OM1 5 5 A06 S11
OM2 3 3 A98 S01
OF1 3 3 A07 S10
OF2 3 2 A06 A07 A08 S09
OF3 2 1 A98 A98 A04 S05
OM3 2 1 A96 S97 A00 A00
OM4 2 1 (1)
a
A98 S00
OF4 1 (1)
b
1 A96 S97 A98 A98
Total 21 17
Harrier MHM1 7 6 A06 A12
MHF1 5 5 A07 S12
MHF2 5 4 A05 A09
MHM2 2 1 A06 A07
MHF3 1 (1)
c
1 A04 A05
MHM3 1 (1)
c
1 A07 A08
Total 21 18
O: osprey; MH: marsh harrier; M: male; F: female; A: autumn; S: spring. In four
cases where the transmitter failed, the bird was fitted with a new transmitter,
providing a second tracking period. Incomplete tracks are in parentheses.
a
Sahara crossing only.
b
Sahara crossing included.
c
Europe crossing only.
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187178
as an index of duration in our analyses (this index thus corresponds
to the inverse of mean latitudinal speed of migration).
Repeatability is an estimate of the fractionof total variance (sum
of between- and within-individual variation) that refers to
between-individual variation (Nakagawa &Schielzeth, 2010). This
means that, when comparing two cases where the level of within-
individual variation is the same, repeatability will be highest for the
case with largest between-individual variation. Comparing cases
with the same between-individual variation, repeatability will of
course be highest for the case with lowest within-individual vari-
ation. To disentangle the relative contribution of the two sources of
the variation between the species, we compared the within- and
between-individual variances using Fisher's Fratios significance
test (Sokal &Rohlf, 1995). Between-individual variance was calcu-
lated from the mean values of the ndifferent individuals, with
degrees of freedom equal to (n1). Within-individual variance
was calculated as the mean of the variances of the nindividuals,
with degrees of freedom equal to (n1).
Finally, we identified and analysed apparent intermediate goal
areas (i.e. geographically limitedareas that were used for stopover or
passed in transit during repeated journeys) of the four ospreys and
the three harriers with at least two full round trips. These journeys
were recorded with transmitters typically providing good-quality
positional data on a daily basis, allowing us to delimit stopover
sites. An area was considered to be a stopover site if a bird moved
less than 50 km/day along the seasonally appropriate migratory
course. An apparent goal area was subsequently defined as an area
where an individual made atleast (1) one stopover and one passage
(defined as a position during active migratory flight) or (2) three
passages, within a radius of 50 km. We considered these criteria as
an indication that an area was not revisited just by chance but
formed a key element of the individual's migration routine.
After identifying these areas for each individual (Tables A1, A2),
we calculated the ‘recurrence’(dimensionless) at each apparent
goal area (the proportion of journeys when the given area had been
visited by the relevant individual, including both stopover and
passage visits) in total (seasons combined), in autumn only, and in
spring only. We also determined the ‘seasonal overlap’(dimen-
sionless) in the use of a given area, by multiplying the autumn and
spring recurrence estimates (i.e. the overlap corresponding to the
average probability of visits to a given area during both seasons of
the same annual migration cycle).
To use these proportions to test for possible differences between
the two species, we first normalized our variables following the
suggested methodology by Warton and Hui (2010). To do that, we
used the logit of the raw values, log(p/(1 p)), where pis the un-
transformed recurrence, adding the minimum nonzero proportion
of the sample to both the numerator and denominator of the logit
function to solve the problem of sample proportions that were
equal to 0 and 1. To explore whether recurrence and overlap differed
between the species, we ran linear mixed models with the logit-
transformed values of autumn, spring and total site use recur-
rence at apparent goal areas, as well as the seasonal overlap in site
goal area visits as the dependent variable, and with species as a
fixed factor and individual as a random factor. Tests were performed
using the maximum likelihood ratio method. All analyses were
performed in IBM SPSS Statistics, v. 21 (IBM, Armonk, NY, U.S.A.).
Ethical Note
We captured birds as adults at their nesting sites when their
chicks were at least 3 weeks old (marsh harriers) or 5 weeks old
(ospreys). Trapping methods are described by Alerstam et al. (2006)
and Strandberg et al. (2008). We attached the transmitters as
(b) (c)
(d) (e) (f)
(a)
46o
36o
26o
46o
36o
26o
Figure 1. Maps showing the routes of eight adult ospreys (first row) and six adult marsh harriers (second row) that completed at least one round trip between the breeding grounds
in Sweden and the wintering quarters in West Africa during 1996e2012. Each panel highlights the three individuals with most repeated journeys of each species (a: OM1; b: OM2;
c: OF1, d: MHM1; e: MHF1, f: MHF2; see Table 1 for details) in blue (autumn) and red (spring), as well as the trips of all other individuals of the species in grey.
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187 179
backpack harnesses and released the birds near the nest within 1 h
of capture. The devices weighed 30e35 g for ospreys (1.5e2.5% of
the birds' body mass) and 18e22 g (2.7e3.3%) for marsh harriers
(Microwave Telemetry, Inc.). The capture did not cause any
increased nest failure beyond that expected in nests without inter-
vention, as no osprey and two (out of six) marsh harriers of the study
failed to breed during the year of tagging, which is lower (ospreys,
28%) or within (marsh harriers, 24e35%) the range of nest desertion
known for nontagged birds (Cramp &Simmons, 2004; Sternalski,
Blanc, Augiron, Rocheteau, &Bretagnolle, 2013). The temporal and
spatial patterns of mortality throughout the annual cycle have
recently been analysed for these species (Klaassen et al., 2014). The
estimated mean annual survival of the satellite-tracked birds was
somewhat lower (marsh harriers/ospreys: 0.5/0.6; N¼17/18) than
the expected annual survival rate (0.7/0.8), indicating that the
transmitters may possibly have had a small adverse effect on the
long-term survival of the birds or that mortality was slightly over-
estimated because some cases of probable deaths may in fact have
been due to transmitter failures (see more detailed discussion in
Klaassen et al., 2014). The meta-analysis by Barron, Brawn, and
Weatherhead (2010) also indicated the lack of a clear effect of cap-
ture and transmitter attachment on nest success and survival rate of
birds. Captures of the raptors and the use of radiotransmitters were
approved by the Ethical Committee in Malm€
o-Lund (permissions
M204-06 and M27-10).
RESULTS
Differences Between Individuals
Individual osprey showed highly significant differences in their
routes (longitude) during both autumn and spring migration
(except at 26
N in spring; Table 2). In contrast, routes of individual
marsh harriers did not differ significantly (except at 36
Nin
autumn). Individual marsh harriers, but not ospreys, showed highly
significant differences in timing of migration (except at 36
Nin
spring). For duration of migration, we found highly significant
differences between individual marsh harriers during both autumn
and spring, and also for individual ospreys during autumn
(although not as pronounced as for the marsh harrier) but not
during spring (Table 2). These patterns suggest contrasting indi-
vidual consistency in routes and timing of migration for these two
species.
Individual Repeatability
A strong effect of individual will be reflected by a high repeat-
ability estimate approaching unity, while a lack of an effect of in-
dividual will be associated with close to zero repeatability. Hence,
repeatability in route was often high and approaching unity for
ospreys but close to zero for the marsh harrier, and vice versa for
timing and duration (Table 3). However, the 95% confidence in-
tervals (CIs) were often wide. Still, in all but one case (across the
Sahara in spring), route repeatability was significant (exceeding
zero) for ospreys, but in no case was it significant for the marsh
harrier. Furthermore, in two cases there was no overlap in CIs for
route repeatability in ospreys and marsh harriers, suggesting sig-
nificant species differences. On the other hand, CIs for the repeat-
ability in timing were generally overlapping zero in all but one case
(marsh harrier crossing the Sahara in spring). Repeatability in
duration was clearly significant for the marsh harrier during
autumn migration (Table 3).
We also compared the variance at the within- and between-
individual levels between species (Table A3). There were no sig-
nificant species differences in migratory timing, while between-
individual variance in duration was significantly larger for marsh
harriers. The between-individual variance in route (longitude) was
significantly larger for ospreys than for marsh harriers (except at
36
N in autumn). In Europe, ospreys also showed a significantly
lower within-individual variance in route (Table A3).
Goal Area Fidelity
Four ospreys and three marsh harriers were tracked for at least
two full annual migrations, and for these individuals we identified
four to seven potential intermediate goal areas for each individual
(Tables A1, A2). The importance of these areas can be inferred by
the frequency of visits; however, there were consistent differences
between the two raptor species (Fig. 2). There were proportionally
many fewer instances of no visits (represented as white sections in
Fig. 2) for these sites among ospreys than among marsh harriers.
Also, remarkably, all four ospreys visited at least one site on all
journeys (autumn and spring), whereas we found no such consis-
tently visited sites among any of the marsh harriers, which in turn
tended to use season-specific stopover sites. Ospreys also used
fewer stopover sites (four or five) than marsh harriers (six or seven
sites). However, marsh harriers used potential intermediate goal
areas more facultatively and readily skipped a given area in a
certain year or even alternated among them during different years
of the study period (see all instances of ‘no visit’in all possible
intermediate goal areas in Fig. 2b).
The total recurrence of visits to apparent goal areas was on
average almost twice as high for ospreys (mean ±SD ¼0.69 ±0.23)
than for marsh harriers (0.37 ±0.16). Furthermore, the seasonal
overlap (corresponding to the probability of visiting these sites in
both migratory seasons in a given year) was on average rather high
for the osprey (0.46 ±0.37) but low for the marsh harrier
(0.09 ±0.13; Fig. 3). Logit-transformed recurrence and overlap
values showed significant differences between species in autumn
(F
1,5.4
¼12.6, N¼2, P¼0.02) and spring (F
1,34
¼4.8, N¼2,
Table 2
ANOVA tests of the effect of individual on route, timing and duration of migration for ospreys and marsh harriers
Season Latitude Longitude Date Duration
Osprey Harrier Osprey Harrier Osprey Harrier
Autumn 46
N 68.9
***(8,22)
1.6
(6,23)
2.9
*(8,22)
5.9
**(6,23)
5.5
**(7,20)
82.6
***(4,19)
36
N 58.1
***(7,20)
5.8
**(4,19)
2.9
(7,20)
9.4
**(4,19)
26
N 11.0
***(7,20)
1.7
(4,19)
1.4
(7,20)
11.1
***(4,19)
Spring 26
N 1.2
(4,13)
0.3
(3,15)
3.8
(4,13)
22.1
***(3,15)
2.7
(4,13)
15.3
**(3,15)
36
N 21.5
***(4,13)
0.5
(3,15)
3.7
(4,13)
3.3
(3,15)
46
N 492.9
***(4,13)
0.7
(3,15)
1.3
(4,13)
9.5
**(3,15)
We analysed variation between individuals in route (longitude) and timing (date) at three latitudes, corresponding to regions in Europe (46
N), the Mediterranean (36
N) and
the Sahara (26
N), during autumn and spring migration. Fvalues, sample sizes (number of individuals, total number of observations) and significance levels (*P<0.05;
**P<0.01; ***P<0.001) are given for each one-way ANOVA test (numerator and denominator degrees of freedom for Fvalues are the number of individuals 1 and the
number of observations 1, respectively).
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187180
P¼0.04). Also, total recurrence (F
1,36
¼25.2, N¼2, P<0.001) and
seasonal overlap (F
1,36
¼16.9, N¼2, P<0.001) differed signifi-
cantly between species.
DISCUSSION
We found a consistent pattern in individual route repeatability
(as reflected by longitude values) of high and clearly significant
estimates for ospreys but low and statistically nonsignificant esti-
mates for marsh harriers (see also Vardanis et al., 2011, for marsh
harriers). Similarly, we found the opposite species-specific pattern
for timing and duration of migration, but this effect was estimated
with higher uncertainty (CIs in Table 3). Our aim in the present
study was to determine whether the repeatability patterns in
migratory space and time across the entire annual cycle differ be-
tween ospreys and marsh harriers. We evaluated, for the first time,
the long-term, individual-based migration patterns of these two
species using high-resolution satellite-tracking data. However, we
are well aware that our sample sizes are rather limited for precise
repeatability estimates (cf. Wolak, Fairbairn, &Paulsen, 2012), as
the often wide 95% CIs of Table 3 suggest. Thus, our specific
repeatability estimates need to be interpreted with caution.
The Geographical Dimension: Repeatability in Route and Stopover
Site Use
We suggest that the difference in route repeatability between
the two species is mainly explained by two factors: (1) larger
between-individual variation in routes in ospreys compared to
marsh harriers (Fig. 1) and (2) a higher degree of recurrently used
stopover sites by individual ospreys compared to marsh harriers
(Fig. 3).
The within-individual variation in routes was indeed rather
similar in the two raptor species (for example, average within-
individual standard deviation at 36
N was 314 km for the marsh
harrier and 182 km for the osprey) compared to the standard de-
viation in routes between individuals, which was clearly larger for
ospreys (730 km) than marsh harriers (253 km; Table A3,Fig. 1),
Table 3
Repeatability estimates (r) for routes, timing and duration of autumn and spring migration of ospreys and marsh harriers
Season Latitude Longitude Date Duration
Osprey Harrier Osprey Harrier Osprey Harrier
Autumn 46
N 0.88 0.06 0.17 0.35 0.35 0.92
0.71e1.05 0.22e0.34 0.16e0.5 0.16e0.86 0.12e0.82 0.72e1.13
36
N 0.87 0.41 0.19 0.55
0.68e1.07 0.42e1.25 0.21e0.58 0.24e1.34
26
N 0.54 0.09 0.04 0.60
0.08e1.01 0.49e0.67 0.21e0.3 0.16e1.35
Spring 26
N 0.04 0.15 0.38 0.81 0.26 0.74
0.67e0.75 0.59e0.28 0.58e1.33 0.06e1.56 0.67e1.19 0.21e1.69
36
N 0.81 011 0.37 0.32
0.34e1.29 0.71e0.50 0.58e1.32 1.13e1.77
46
N 0.99 0.13 0.07 0.63
0.96e1.02 0.81e0.67 0.68e0.81 0.57e1.84
We calculated repeatability in route (longitude) and timing (date) at three latitudes, corresponding to regions in Europe (46
N), the Mediterranean (36
N) and the Sahara
(26
N). Sample sizes for each estimate are given in Table 2. The 95% confidence intervals (calculated from Fratios according to Nakagawa &Schielzeth, 2010) are given below
each estimate. Intervals that did not overlap with zero are shown in bold.
Number of migratory journeys
Individuals
OM1 OM2 OF1 OF2 MHM1 MHF1 MHF2
0
2
4
6
8
10
12
14 (a) S-no visit
S-passage
S-stopover
A-no visit
A-passage
A-stopover
(b)
Figure 2. Repeated site use in (a) four ospreys and (b) three marsh harriers that were tracked for at least 2 complete years. The seven individuals are arranged along the Xaxis, each
with four to six bars representing the apparent goal areas identified for that individual (see also Table 1 for details of trips and Tables A1, A2 for details of the goal areas). Hence, for
osprey OM1, there were four apparent goal areas, and for osprey OM2, there were five apparent goal areas (Tables A1, A2), and so on. The Yaxis shows the number of migratory
journeys, with data for spring migration (S) in grey and autumn migration (A) in black. Filled parts of bars indicate journeys when the area was used for stopover, hatched parts of
bars indicate when the area was briefly visited (passage) while open parts of bars denote journeys when the site was not visited by the bird.
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187 181
reflecting broad-front migration in ospreys and a narrow migration
corridor in marsh harriers. As a consequence, route repeatability
should be high for ospreys because the route of one individual
osprey is unlikely to be similar to the route of another osprey. In
contrast, the routes of individual marsh harriers were all in a rather
narrow migration corridor, with between-individual variation
small compared with the within-individual variation in route, and
repeatability was thus closer to zero. In addition, the ospreys' habit
of revisiting individual-specific goal areas (often used for stopover;
already described by Alerstam et al., 2006) meant that routes of the
same osprey in different years converged towards these apparent
goal areas and this contributed to the overall degree of repeatability
in routes for this species.
Why is there a broad geographical migration front of ospreys
(i.e. large variation in routes between individuals and between
siblings from the same broods; cf. €
Osterl€
of, 1977) and a high degree
of fidelity to individual-specific stopover areas? One possibility
may be related to ospreys' reliance on fish for food. High-quality
foraging sites are probably relatively scarce, but widespread
throughout Europe, which could promote broad-front migration.
Furthermore, hunting efficiency is possibly reduced with increasing
densities of foraging ospreys, promoting thinning among in-
dividuals and thus wider variation in routes (cf. Alerstam, 1990).
Hunting efficiency may improve with increasing local knowledge
about the area, and thus, would explain high fidelity to specific
stopover sites. Another reason for high stopover site fidelity could
be that the quality of these sites, as reflected by the availability of a
largely renewable food source, may show little fluctuation between
seasons and years.
Among solitary migrants, the marsh harrier and various small
songbirds seem to show a lower degree of route repeatability and
nonbreeding site fidelity (Catry et al., 2004; Drost, 1941; Stanley
et al., 2012; Vardanis et al., 2011) than the osprey. Although
marsh harriers and songbirds differ in many aspects (e.g. size and
flight mode), they are similar in that their potential feeding habitat
is much more widespread and prey availability may be spatially
more variable compared to a feeding specialist like the osprey.
Marsh harriers probably have more to gain from being flexible in
their choice of stopover sites, meaning that the relative role of other
factors potentially affecting route choice, such as seasonal variation
in the distribution of favourable foraging areas and seasonal vari-
ation in wind, becomes more important.
The Temporal Dimension: Repeatability in Timing and Duration
Even though repeatability in timing and duration of migration
was consistently larger for the marsh harrier than the osprey, the
confidence intervals (Table 3) indicate that this tendency is more
suggestive rather than conclusive compared to the geographical
dimension discussed above. The timing of migration is presumably
determined by an interaction between (1) endogenous factors and
developmental processes associated with the individuals' annual
cycles of migration, breeding and moult and (2) environmental
factors associated with variability of resources and travel condi-
tions. However, it remains unclear why ospreys tend to have a
larger individual plasticity in timing and duration of migration than
marsh harriers. One interesting possibility is that the potential
foraging constraints that make ospreys more repeatable with
respect to their routes and use of individual-specific stopover sites
may influence migratory timing in the opposite way, for example if
stopover duration is related to the foraging conditions at a given
site in a given year, thus promoting relatively high flexibility in
timing between years.
It is somewhat paradoxical that marsh harriers, despite their
high individual consistency in annual timing of migration, have
advanced their autumn passage through southern France (i.e.
including the Swedish population) twice as much as ospreys have
over the last 30 years (Filippi-Codaccioni et al., 2010). While the
mechanisms driving this population-level advance in a species with
little individual plasticity remains elusive, this pattern has also
been observed among Icelandic black-tailed godwits, Limosa limosa
islandica (Gill et al., 2014) and migratory black kites, Milvus migrans
(Sergio et al., 2014).
Timing performance may also change with experience and
learning. Sergio et al. (2014) found that the most dramatic indi-
vidual ‘improvement’in migratory timing occurs among young
prebreeding birds. However, in our study we included only adult
birds, and we assumed no difference in experience between birds
within the two species. An additional possible limitation could be
due to the differential timing (Alerstam et al., 2006) and duration
(Strandberg et al., 2008) of migration between the sexes. However,
the small sample size precludes testing for sex effects, but the
balanced sex ratio in our study makes such a potential effect less
likely to bias our results.
Individual Consistency Versus Flexibility in Different Migratory
Traits
A considerable number of repeatability studies have provided
evidence supporting significant consistency in the timing of bird
migration among a variety of species (Conklin et al., 2013; Gill et al.,
2014; Lopez-Lopez et al., 2014; Lourenço et al., 2011; Pulido, 2007b;
Thorup et al., 2013), whereas evidence for relatively low and some-
times nonsignificant repeatability in routes has been reported for
three raptor species (marsh harrier: Vardanis et al., 2011;Egyptian
vulture, Neophron percnopterus:Lopez-Lopez et al., 2014; black kite:
Sergio et al., 2014), one songbird (Stanley et a l., 2012)andonepelagic
long-distance migrant (Dias et al., 2013). Even though it may be
difficult to compare repeatability estimates across studies because of
variation in measurement precision, the present set of published
Seasonal
overlap
Total
recurrence
Spring
recurrence
Autumn
recurrence
Mean rate ± SD
1
0.75
0.5
0.25
0
Marsh harrier
Osprey
Figure 3. Mean ±SD rates of autumn, spring and total recurrence at apparent goal
areas, as well as the seasonal overlap in goal area visits, for the four ospreys (N¼19
goal areas used) and the three marsh harriers (N¼19 goal areas used) that were
tracked for at least 2 complete years (based on the data of repeated goal area use
shown in Fig. 2).
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187182
studies would make it tempting to suggest that a higher repeatability
in timing compared to routes may be a general feature in bird
migration. This may be a consequence of constraints set by genetic/
endogenous mechanisms controlling the timing of bird migration,
while advanced navigation capabilities allow flexibility in routes
(associated with environmental variation inwind and resources), also
for specieswith a rather strict fidelity to breeding and wintering sites.
However, the case of the osprey shows a contrasting pattern
with relatively lower repeatability in timing and higher repeat-
ability estimates in routes. Thus, the relative level of individual
variation present in different time- and space-associated behav-
ioural traits across avian migrants may not reflect general adapta-
tions or constraints in long-distance migration common to all long-
distance migrants (Piersma, P
erez-Tris, Mouritsen, Bauchinger, &
Bairlein, 2005), but might, instead, be shaped by species-specific
differences in life history and ecological characteristics, as well as
the importance of selective forces (time, energy and safety).
Future Directions
Estimating repeatability is a tool to facilitate inferences about
the evolutionary ecology of behavioural traits, and, thus, its esti-
mation may not be an aim in itself. Despite the large heterogeneity
of repeatability studies in the bird migration literature, compari-
sons of restimates between different groups of birds (e.g. sexes,
populations, species within and between different migratory sys-
tems, as well as different stages of the annual cycle) hold great
potential in furthering our understanding of the adaptive value of
individuality in migratory behaviour. Comparisons across species
and annual stages may, for example, identify suitable traits ac-
commodating genetic variation and, thus, possibly larger potential
for evolutionary change (van Noordwijk et al., 2006). Furthermore,
such an approach, in combination with additional lines of data (e.g.
stopover behaviour, mortality patterns, etc.) can help determine
the relevant ecological conditions (Pulido, 2007b) and optimization
factors (e.g. time, energy, mortality risk; Alerstam, 2011) that shape
different migratory strategies. Finally, future repeatability analyses
of individual migration histories, including the initial journeys in
the life of an individual (McKinnon, Fraser, Stanley, &Stutchbury,
2014; Schiffner, Pavkovic, Siegmund, &Wiltschko, 2011; Sergio
et al., 2014; Stout, Greene, &Postupalsky, 2009), will be needed
in order to disentangle the relative importance and interaction
between genetic and environmentally induced responses, as well
as experience-based learning and cultural and social influences on
the observed strategies in migratory vertebrates.
Acknowledgments
We thank all collaborators for their help, especially Mikael Hake,
Nils Kjell
en and Patrik Olofsson. We are grateful for support and
excellent tracking equipment from Microwave Telemetry Inc. We
acknowledge the use of the Maptool program for graphics in this
paper (www.seaturtle.org). Tracking data are stored in Movebank
(movebank.org). This work was supported by grants from the
Swedish Research Council (621-2009-3586 and 621-2012-3221 to
T.A.) and from Kungliga Fysiografiska S€
allskapet in Lund, Sweden
(to R.H.G.K.). We are very grateful for many valuable and
constructive comments on the manuscript by Jenny Gill and two
anonymous referees.
References
Alerstam, T. (1990). Bird migration. Cambridge, U.K.: Cambridge University Press.
Alerstam, T. (2006). Conflicting evidence about long-distance animal navigation.
Science, 313(5788), 791e794. http://dx.doi.org/10.1126/science.1129048.
Alerstam, T. (2011). Optimal bird migration revisited. Journal of Ornithology, 152(1),
5e23. http://dx.doi.org/10.1007/s10336-011-0694-1.
Alerstam, T., Hake, M., & Kjell
en, N. (2006). Temporal and spatial patterns of
repeated migratory journeys by ospreys. Animal Behaviour, 71, 555e566. http://
dx.doi.org/10.1016/j.anbehav.2005.05.016.
Baker, R. R. (1978). The evolutionary ecology of animal migration. London, U.K.:
Hodder &Stoughton.
Barron, D. G., Brawn, J. D., & Weatherhead, P. J. (2010). Meta-analysis of transmitter
effects on avian behaviour and ecology. Methods in Ecology and Evolution, 1(2),
180 e187.
Bell, A. M., Hankison, S. J., & Laskowski, K. L. (2009). The repeatability of behaviour:
a meta-analysis. Animal Behaviour, 77(4), 771e783. http://dx.doi.org/10.1016/
j.anbehav.2008.12.022.
Broderick, A. C., Coyne, M. S., Fuller, W. J., Glen, F., & Godley, B. J. (2007). Fidelity and
over-wintering of sea turtles. Proceedings of the Royal Society B: Biological Sci-
ences, 274(1617), 1533e1539. http://dx.doi.org/10.1098/rspb.2007.0211.
Catry, P., Encarnaç~
ao, V., Araújo, A., Fearon, P., Fearon, A., Armelin, M., et al. (2004).
Are long-distance migrant passerines faithful to their stopover sites? Journal of
Avian Biology, 35(2), 170e181. http://dx.doi.org/10.1111/j.0908-
8857.2004.03112.x.
Conklin, J. R., Battley, P. F., & Potter, M. A. (2013). Absolute consistency: individual
versus population variation in annual-cycle schedules of a long-distance
migrant bird. PLoS One, 8(1), e54535. http://dx.doi.org/10.1371/
journal.pone.0054535.
Cramp, S., & Simmons, K. (2004). Birds of the Western Palearctic interactive. Sheffield,
U.K.: BirdGuides.
Dias, M. P., Granadeiro, J. P., & Catry, P. (2013). Individual variability in the migratory
path and stopovers of a long-distance pelagic migrant. Animal Behaviour, 86(2),
359e364. http://dx.doi.org/10.1016/j.anbehav.2013.05.026.
Drost, R. (1941). Zieht der einzelne Vogel stets auf demselben Weg? Ardea, 30,
215e223.
Gill, J. A., Alves, J. A., Sutherland, W. J., Appleton, G. F., Potts, P. M., &
Gunnarsson, T. G. (2014). Why is timing of bird migration advancing when
individuals are not? Proceedings of the Royal Society B: Biological Sciences,
281(1774). http://dx.doi.org/10.1098/rspb.2013.2161.
Filippi-Codaccioni, O., Moussus, J.-P., Urcun, J.-P., & Jiguet, F. (2010). Advanced de-
parture dates in long-distance migratory raptors. Journal of Ornithology, 151(3),
687e694. http://dx.doi.org/10.1007/s10336-010-0500-5.
Fransson, T., & Pettersson, J. (2001). Swedish Bird Ringing Atlas Vol. 1. Divers - Raptors.
Stockholm, Sweden: Swedish Museum of Natural History.
Gudmundsson, G. A., & Alerstam, T. (1998). Optimal map projections for analysing
long-distance migration routes. Journal of Avian Biology, 29(4), 597e605. http://
dx.doi.org/10.2307/3677180.
Hake, M., Kjell
en, N., & Alerstam, T. (2001). Satellite tracking of Swedish os-
preys Pandion haliaetus: autumn migration routes and orientation. Journal
of Avian Biology, 32(1), 47e56. htt p: // dx .doi.org/ 10.1034 /j .160 0 -
048X.2001.320107.x.
Klaassen, R. H. G., Hake, M., Strandberg, R., Koks, B. J., Trierweiler, C., Exo, K.-M., et al.
(2014). When and where does mortality occur in migratory birds? Direct evi-
dence from long-term satellite tracking of raptors. Journal of Animal Ecology,
83(1), 176e184. http://dx.doi.org/10.1111/1365-2656.12135.
Knudsen, E., Lind
en, A., Both, C., Jonz
en, N., Pulido, F., Saino, N., et al. (2011). Chal-
lenging claims in the study of migratory birds and climate change. Biological
Reviews, 86(4), 928e946. http://dx.doi.org/10.1111/j.1469-185X.2011.00179.x.
Lessells, C. M., & Boag, P. T. (1987). Unrepeatable repeatabilities: a commonmistake.
Auk, 104(1), 116e121. http://dx.doi.org/10.2307/4087240.
Lopez-Lopez, P., Garcia-Ripolles, C., & Urios, V. (2014). Individual repeatability in
timing and spatial flexibility of migration routes of trans-Saharan migratory
raptors. Current Zoology, 60(5), 642e652.
Lourenço, P., Kentie, R., Schroeder, J., Groen, N., Hooijmeijer, J., & Piersma, T. (2011).
Repeatable timing of northward departure, arrival and breeding in black-tailed
godwits Limosa l. limosa, but no domino effects. Journal of Ornithology, 152(4),
1023e
1032. http://dx.doi.org/10.1007/s10336-011-0692-3.
Lynch, M., & Walsh, B. (1998). Genetics and analysis of quantitative traits. Sunderland,
MA: Sinauer.
McKinnon, E. A., Fraser, K. C., Stanley, C. Q., & Stutchbury, B. J. M. (2014). Tracking
from the tropics reveals behaviour of juvenile songbirds on their first spring
migration. PLoS One, 9(8), e105605. http://dx.doi.org/10.1371/
journal.pone.0105605.
Mellone, U., Klaassen, R. H. G., García-Ripoll
es, C., Limi~
nana, R., L
opez-L
opez, P.,
Pav
on, D., et al. (2012). Interspecific comparison of the performance of soaring
migrants in relation to morphology, meteorological conditions and migration
strategies. PLoS One, 7(7), e39833. http://dx.doi.org/10.1371/
journal.pone.0039833.
Nakagawa, S., & Schielzeth, H. (2010). Repeatability for Gaussian and non-Gaussian
data: a practical guide for biologists. Biological Reviews, 85(4), 935e956. http://
dx.doi.org/10.1111/j.1469-185X.2010.00141.x.
van Noordwijk,A. J., Pulido, F., Helm,B., Coppack, T.,Delingat, J., Dingle,H., et al. (2006).
A framework for the study of genetic variation in migratory behaviour. Journal of
Ornithology, 147(2), 221e233. http://dx.doi.org/10.1007/s10336-005-0047-z.
€
Osterl€
of, S. (1977). Migration, wintering areas, and site tenacity of the European
osprey Pandion h. haliaetus (L.). Ornis Scandinavica, 8(1), 61e78. http://
dx.doi.org/10.2307/3675988.
Piersma, T., P
erez-Tris, J., Mouritsen, H., Bauchinger, U. L. F., & Bairlein, F. (2005). Is
there a ‘migratory syndrome’common to all migrant birds? Annals of the New
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187 183
York Academy of Sciences, 1046(1), 282e293. http://dx.doi.org/10.1196/
annals.1343.026.
Poole, A. F. (1989). Ospreys: A natural and unnatural history. Cambridge, U.K.:
Cambridge University Press.
Pulido, F. (2007a). The genetics and evolution of avian migration. Bioscience, 57(2),
165e174. http://dx.doi.org/10.1641/b570211.
Pulido, F. (2007b). Phenotypic changes in spring arrival: evolution, phenotypic
plasticity, effects of weather and condition. Climate Research, 35(1e2), 5e23.
http://dx.doi.org/10.3354/cr00711.
Schiffner, I., Pavkovic, T., Siegmund, B., & Wiltschko, R. (2011). Strategies of young
pigeons during ‘map’learning. Journal of Navigation, 64(03), 431e448. http://
dx.doi.org/10.1017/S0373463311000063.
Schofield, G., Hobson, V. J., Fossette, S., Lilley, M. K. S., Katselidis, K. A., & Hays, G. C.
(2010). Fidelity to foraging sites, consistency of migration routes and habitat
modulation of home range by sea turtles. Diversity and Distributions, 16(5),
840e853. http://dx.doi.org/10.1111/j.1472-4642.2010.00694.x.
Sergio, F., Tanferna, A., De Stephanis, R., Jimenez, L. L., Blas, J., Tavecchia, G., et al.
(2014). Individual improvements and selective mortality shape lifelong
migratory performance. Nature, 515(7527), 410e413. http://dx.doi.org/10.1038/
nature13696.
Sokal, R. R., & Rohlf, F. J. (1995). Biometry. New York, NY: W. H. Freeman.
Stanley, C. Q., MacPherson, M., Fraser, K. C., McKinnon, E. A., & Stutchbury, B. J. M.
(2012). Repeat tracking of individual songbirds reveals consistent migration
timing but flexibility in route. PLoS One, 7(7), e40688. http://dx.doi.org/10.1371/
journal.pone.0040688.
Sternalski, A., Blanc, J.-F., Augiron, S., Rocheteau, V., & Bretagnolle, V. (2013).
Comparative breeding performance of marsh harriers Circus aeruginosus along a
gradient of land-use intensification and implications for population manage-
ment. Ibis, 155(1), 55e67. http://dx.doi.org/10.1111/ibi.12003.
Stout, W. E., Greene, V. L., & Postupalsky, S. (2009). Migration routes, reproduction,
and lifespan of a translocated osprey. Wilson Journal of Ornithology, 121(1),
203e206. http://dx.doi.org/10.1676/08-030.1.
Strandberg, R., Klaassen, R. H. G., Hake, M., Olofsson, P., Thorup, K., & Alerstam, T.
(2008). Complex timing of marsh harrier Circus aeruginosus migration due to
pre- and post-migratory movements. Ardea, 96(2), 159e171. http://dx.doi.org/
10.5253/078.096.0202.
Sutherland, W. J. (1998). Evidence for flexibility and constraint in migration sys-
tems. Journal of Avian Biology, 29(4), 441e446. http://dx.doi.org/10.2307/
3677163.
Thorup, K., Vardanis, Y., Tøttrup, A. P., Kristensen, M. W., & Alerstam, T. (2013).
Timing of songbird migration: individual consistency within and between
seasons. Journal of Avian Biology, 44(5), 486e494. http://dx.doi.org/10.1111/
j.1600-048X.2013.05871.x.
Vardanis, Y., Klaassen, R. H. G., Strandberg, R., & Alerstam, T. (2011). Individuality in
bird migration: routes and timing. Biology Letters, 7(4), 502e505. http://
dx.doi.org/10.1098/rsbl.2010.1180.
Warton, D. I., & Hui, F. K. C. (2010). The arcsine is asinine: the analysis of proportions
in ecology. Ecology, 92(1), 3e10. http://dx.doi.org/10.1890/10-0340.1.
Wolak, M. E., Fairbairn, D. J., & Paulsen, Y. R. (2012). Guidelines for estimating
repeatability. Methods in Ecology and Evolution, 3(1), 129e137. http://dx.doi.org/
10.1111/j.2041-210X.2011.00125.x.
Appendix
Table A1
Possible intermediate goal areas (i.e. areas used for repeated stopover or passage) by four ospreys that were tracked for at least 2 complete years
Individual Country (coordinates) Autumn Spring
Year Use Duration (days) Year Use Duration (days)
OM1 1. France (47
42
0
N, 6
35
0
E) 2006 SO 15 2007 P
2007 SO 30 2008 SO 5
2008 SO 31 2009 SO 5
2009 SO 31 2010 SO 5
2010 SO 22 2011 SO 6
2. France (44
23
0
N, 3
12
0
E) 2006 e2007 P
2007 P 2008 e
2008 e2009 e
2009 e2010 P
2010 SO 2 2011 P
3. Algeria (29
11
0
N, 7
12
0
W) 2006 P 2007 e
2007 P 2008 e
2008 P 2009 e
2009 N 2010 e
2010 e2011 e
4. Mauritania (19
36
0
N, 13
5
0
W) 2006 e2007 P
2007 P 2008 P
2008 P 2009 P
2009 P 2010 e
2010 P 2011 P
OM2 1. Poland (53
6
0
N, 16
11
0
E) 1998 P 1999 e
1999 e2000 e
2000 SO 2 2001 e
2. Poland (52
36
0
N, 14
18
0
E) 1998 e1999 SO 5
1999 P 2000 e
2000 e2001 P
3. Poland (51
N, 15
18
0
E) 1998 SO 16 1999 P
1999 SO 10 2000 SO 1
2000 SO 8 2001 SO 2
4. Italy (45
30
0
N, 13
18
0
E) 1998 P 1999 P
1999 e2000 e
2000 P 2001 P
5. Tunysia (35
17
0
N, 10
24
0
E) 1998 SO 8 1999 e
1999 P 2000 e
2000 P 2001 e
OF1 1. Germany (51
23
0
N, 10
54
0
E) 2007 SO 16 2008 SO 11
2008 SO 20 2009 SO 6
2009 SO 34 2010 P
2. W. Sahara (25
11
0
N, 11
11
0
W) 2007 e2008 P
2008 P 2009 e
2009 P 2010 P
3. France (44
17
0
N, 0
30
0
E) 2007 P 2008 e
2008 P 2009 e
2009 P 2010 x
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187184
Table A1 (continued )
Individual Country (coordinates) Autumn Spring
Year Use Duration (days) Year Use Duration (days)
4. Spain (36
12
0
N, 5
24
0
W) 2007 P 2008 P
2008 P 2009 P
2009 P 2010 x
5. Morocco (34
6
0
N, 6
17
0
W) 2007 SO 5 2008 e
2008 SO 8 2009 P
2009 SO 3 2010 x
OF2 1. Morocco (33
12
0
N, 3
36
0
W) 2006 e2007 e
2007 P 2008 x
2008 P 2009 P
2. Spain (36
23
0
N, 5
17
0
W) 2006 SO 6 2007 P
2007 e2008 x
2008 SO 10 2009 e
3. Germany (51
17
0
N, 10
35
0
E) 2006 P 2007 P
2007 e2008 e
2008 P 2009 x
4. Spain (41
30
0
N, 1
6
0
W) 2006 P 2007 P
2007 P 2008 x
2008 e2009 P
5. Germany (50
17
0
N, 12
18
0
E) 2006 SO 4 2007 P
2007 SO 44 2008 x
2008 SO 24 2009 x
SO: stopover; P: passage; e: absence; x: data gap. We used the following criteria as minimum requirements to define an area as a possible goal: (1) one stopover and one
passage, (2) two stopovers or (3) three passages (within a radius of 50 km). Duration of stopover is also shown.
Table A2
Possible intermediate goal areas (i.e. areas used for repeated stopover or passage) by three marsh harriers that were tracked for at least 2 complete years
Individual Area (coordinates) Autumn Spring
Year Use Duration (days) Year Use Duration (days)
MHM1 1. Spain (40
6
0
N, 1
18
0
W) 2006 SO 3 2007 e
2007 e2008 P
2008 e2009 e
2009 e2010 e
2010 e2011 e
2011 e2012 e
2012 e
2. Spain (39
N, 2
W) 2006 P 2007 e
2007 P 2008 e
2008 e2009 e
2009 e2010 e
2010 e2011 e
2011 SO 1 2012 e
2012 e
3. Spain (36
36
0
N, 5
54
0
W) 2006 e2007 e
2007 e2008 e
2008 e2009 e
2009 e2010 P
2010 e2011 P
2011 e2012 SO 16
2012 e
4. Morocco (32
12
0
N, 7
24
0
W) 2006 e2007 SO 27
2007 e2008 SO 22
2008 e2009 SO 20
2009 e2010 SO 42
2010 e2011 e
2011 e2012 P
2012 e
5. Algeria (31
30
0
e28
30
0
N, 4
W)
a
2006 P 2007 e
2007 SO 1 2008 e
2008 P 2009 e
2009 P 2010 e
2010 P 2011 P
2011 P 2012 e
2012 SO 1
6. Mauritania (13
30
0
N, 20
18
0
W) 2006 e2007 P
2007 e2008 e
2008 e2009 e
2009 e2010 e
2010 e2011 SO 12
2011 e2012 e
(continued on next page)
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187 185
Table A2 (continued )
Individual Area (coordinates) Autumn Spring
Year Use Duration (days) Year Use Duration (days)
2012 e
MHF1 1. Germany (52
23
0
N, 10
18
0
E) 2007 e2008 SO 1
2008 e2009 e
2009 P 2010 e
2010 e2011 e
2011 e2012 e
2. Germany (51
17
0
N, 8
E) 2007 P 2008 e
2008 e2009 P
2009 P 2010 P
2010 e2011 P
2011 e2012 e
3. France (48
47
0
N, 6
42
0
E) 2007 P 2008 SO 2
2008 e2009 e
2009 e2010 e
2010 e2011 e
2011 e2012 e
4. France (43
12
0
N, 3
23
0
E) 2007 e2008 P
2008 e2009 e
2009 SO 5 2010 P
2010 SO 2 2011 e
2011 P 2012 e
5. Spain (40
36
0
N, 0
36
0
E) 2007 SO 3 2008 SO 1
2008 SO 3 2009 e
2009 P 2010 P
2010 SO 2 2011 e
2011 SO 3 2012 e
6. Spain (38
N, 0
36
0
W) 2007 P 2008 e
2008 P 2009 e
2009 P 2010 e
2010 e2011 e
2011 P 2012 e
7. (Algeria (28
23
0
N, 4
30
0
W) 2007 P 2008 e
2008 P 2009 e
2009 e2010 e
2010 e2011 e
2011 P 2012 e
MHF2 1. Germany (50
e53
N, 9
e11
E) 2005 P 2006 e
2006 e2007 P
2007 e2008 P
2008 e2009 e
2009 P
2. France (46.5
N, 5
E) 2005 e2006 P
2006 SO 1 2007 e
2007 S 3 2008 P
2008 e2009 e
2009 P
3. Spain (36.6
N, 5.7
W) 2005 e2006 e
2006 e2007 SO 6
2007 e2008 SO 5
2008 e2009 SO 4
2009 e
4. Algeria (34.5
N, 0
) 2005 P 2006 e
2006 P 2007 e
2007 P 2008 e
2008 P 2009 e
2009 e
5. Algeria (30
N, 2.5
W) 2005 P 2006 e
2006 P 2007 e
2007 P 2008 e
2008 P 2009 e
2009 e
6. Morocco (29.7
N, 9.7
W) 2005 e2006 P
2006 e2007 P
2007 e2008 e
2008 e2009 P
2009 e
SO: stopover; P: passage; e: absence. We used the following criteria as minimum requirements to define an area as a possible goal: (1) one stopover and one passage, (2) two
stopovers or (3) three passages (within a radius of 50 km). Duration of stopover is also shown.
a
Possibly identical flight paths.
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187186
Table A3
Within- and between-individual variances (s
2
), as well as Fratios for routes, timing and duration of migration of ospreys and marsh harriers
Season Latitude Species df Longitude Date Duration
Within Between Within Between Within Between
Autumn 46
N Osprey 7 1.90 36.36 73.93 121.28
Harrier 5 14.15 3.96 49.41 52.61
7.46 9.19 1.50 2.31
36
N Osprey 6 4.10 65.91 98.83 161.43 0.07 0.16
Harrier 3 12.21 7.95 84.06 128.65 0.06 1.47
2.98 8.29 1.18 0.80 1.14 9.45
26
N Osprey 6 12.15 27.83 202.17 162.81
Harrier 3 9.15 2.48 82.98 143.34
1.33 11.24 2.44 1.14
Spring 26
N Osprey 3 21.90 6.55 36.47 63.08
Harrier 2 3.16 0.24 56.49 279.54
6.93 27.56 1.55 4.43
36
N Osprey 3 2.78 17.81 36.95 60.02 0.18 0.12
Harrier 2 1.37 0.13 44.60 34.37 0.04 0.11
2.02 133.34 1.21 1.75 4.81 1.12
46
N Osprey 3 0.21 33.30 56.04 32.89
Harrier 2 2.22 0.35 17.93 30.24
10.75 95.02 3.13 1.09
Variances for route (degrees of longitude) and timing (date) were calculated for three latitudes, corresponding to regions in Europe (46
N), the Mediterranean (36
N) and the
Sahara (26
N). Index of duration refers to days per degree of latitude (see Methods). Degrees of freedom (df ¼number of individuals 1) are provided for each estimate.
Differences between species in variances were tested by Fisher's Fratios (shown in italics); statistical significance based on an Fdistribution table for P¼0.05 are shown in
bold.
Y. Vardanis et al. / Animal Behaviour 113 (2016) 177e187 187
