Page 56 - Panuccio2012
P. 56
Differential migration in Western Marsh Harrier
a consequence also the wintering grounds of this 2004, Strandberg et al. 2008, Agostini and
species are extensively distributed: birds breeding Panuccio 2010). A Generalized Linear Model
in Sweden winter in Africa south of Sahara, while (GLM) with negative binomial distribution was
individuals belonging to western France are mostly applied using the “MASS” package included in R
residents (Sternalski et al. 2008, Strandberg et al. software (Venables and Riley 2002). Latitude and
2008). longitude as well as age class (juvenile, adult) were
Brown et al. (1982) supposed that European used as independent factors to explain variation in
populations show a leap-frog migration, but no the loxodrome distance (km) between the ringing
data were provided to support this statement. site (breeding area or birthplace) and the recovery
Some recent papers investigated the movements (wintering site). In the analysis the age of birds
of population of limited geographical areas with was defined according to the recovery data, and
visual observations, traditional radio-telemetry therefore a bird ringed at the nest and recovered
and, more recently, by satellite tracking (Sternalski at least in its second or following winter was
et al. 2008, Strandberg et al. 2008, Klaassen et considered an adult. We did not take into account
al. 2010b, Agostini and Panuccio 2010). Here the sex category due to the low sample of sexed
we analyze migration patterns of Western Marsh birds.
Harriers using ring recoveries data from a large Age related variation concerning the location
portion of Europe. A previous study compared of wintering sub-Saharan areas was investigated
results of ringing recoveries data with results using Mann-Whitney U-test to compare longitude
provided by satellite tracking of some species of wintering sites. V-test was used to compare the
of raptors, among which the Western Marsh migratory directions being a circular statistic test
Harrier (Strandberg et al. 2009). It suggested that allows to verify if two migratory directions
that the interpretation of the temporal patterns are similarly oriented. Moreover we calculated
from recovery data is problematic but on the the dispersal index (r) that in circular statistic is
other hand, ring recoveries and satellite tracking the length of the mean vector (Fowler and Cohen
provide more or less the same picture concerning 1992). Finally, ArcGIS 9.1 software was used to
the geographical pattern of data distribution. project and visually inspect the data.
Moreover ring recoveries allow to manage a higher Results
amount of data than satellite tracking and, as a Our GLM shows a significant, positive,
result this method allows to reduce the bias due to relationship (Tab.1) between the length of the
a small sample size. migratory journey and both latitude (Fig. 1;
Our aim was to compare the distances between P<0.01) and longitude (P<0.01) of the breeding
breeding and wintering grounds in relation to sites but not with age class (P>0.05), showing that
longitude, latitude and age class. individuals belonging to northerly and easterly
Materials and Methods populations migrate farther than individuals
Ringing data of Western Marsh Harriers were breeding in Western and Southern Europe.
obtained from the EURING database. In the 43 individuals were recovered in Africa south of
analysis were considered only high accurate data Sahara Desert, among them 21 were juveniles and
concerning coordinates of capture and recovery 22 adults (48.8% vs. 51.2%; χ²= 0.02, d.f.=1,
with a maximum error of 50 km (EURING n.s.), comparing longitudes of breeding sites of
categories from 0 to 4). After filtering, we retained these juveniles and adults no differences were
for the analyses a total of 345 individuals ringed found (Mann-Whitney U-test; U=183, n.s.).
during the breeding season (10 may - 10 July) On the other hand, comparing the longitude of
and recovered during winter (15 October - 28 wintering sites in sub-Saharan Africa, juveniles
February), selecting these periods according to wintered significantly western than adults (Mann-
the migratory phenology of the species (Génsbøl Whitney U-test; U=134, P<0.05).
56