Survival and dispersal in a new Barnacle Goose population 75
©Wildfowl & Wetlands Trust
Wildfowl
(2013) 63: 72–89
Annual re-sighting rates of marked geese
in the Netherlands are generally very high and
typically exceed 90%. Hence, it is rare for
birds that are alive to go undetected for long
periods of time. All observations made after
a bird had not been observed for at least two
years were checked. Usually, these concerned
single observations after which the bird was
never seen again, they were often entered
by the same observers, and were made at
localities that the birds had never visited
before. These observations were removed.
When more than three observations were
made by
bona fide
observers, or when the
frequency of observations was always low in
a given bird, the observations were retained.
All observations that were made before the
ringing date were removed (66 cases). Using
these criteria, a total of 168 out of 20,860
observations (0.81%) were classed as
incorrect, and 20,692 observations were
retained for analysis.
Counts
Summer counts of all species of geese,
including both adults and young of the year,
were conducted in the Delta area around the
15th of July each year by Sovon, the Dutch
Centre for Field Ornithology, in 2006 and
2007 (van der Jeugd & de Boer 2006; de
Boer & van der Jeugd 2007) and by Centrum
voor Landbouw en Milieu (CLM) in
2007–2011 (Tolkamp & Guldemond 2007;
Guldemond & Tolkamp 2008; Tolkamp &
Guldemond 2009; Visser, Guldemond &
Tolkamp 2010; Den Hollander & Visser
2011). Counts were performed within two
days by a team of volunteers and
professionals and were carried out to
estimate the total numbers of each species
present in each year, to provide data for
planning and evaluating the management of
geese breeding in the Netherlands.
Survival
Survival analyses were performed using
Program MARK (White & Burnham 1999).
Model selection was based on a modified
Akaike Information Criterion (Q-AICc).
Birds ringed as juveniles and birds ringed as
adults were held separately in two groups.
The analyses commenced with a model
where the probability of survival during the
first year after ringing was held separately
from the probability of survival during
subsequent years in both groups, hereafter
referred to as age classes 1 and 2. For birds
ringed as juveniles, survival in age class 1
measures survival during the first year of
life, from fledging until the first summer,
and survival in age class 2 measures survival
during subsequent years. For birds ringed as
adults, survival in both age classes measures
adult survival in birds that are at least one
year of age, although in age class 2, all birds
are one year older than in age class 1. In this
way, we could test for a possible age-effect
on survival for birds ringed as juveniles, and
accommodate a possible transient effect in
adults (Pradel 1997). Re-sightings were
grouped according to calendar year. The
median month of the first observation was
February in each of the years 2005–2009
and 2011, March in 2010 and January in
2012. In all years of the study, 90% of the
first observations of individual geese were
made before the end of May. Hence,
survival was analysed from late winter/
spring until late winter/spring the next
year, with the exception of the first year,
