Content uploaded by Elie Gaget
Author content
All content in this area was uploaded by Elie Gaget on Aug 14, 2018
Content may be subject to copyright.
Long-term decline despite conservation efforts
questions Eurasian Stone-curlew population viability
in intensive farmlands
ELIE GAGET,
1,2,3
* REMI FAY,
1
STEVE AUGIRON,
1,4
ALEXANDRE VILLERS
1,5
&
VINCENT BRETAGNOLLE
1,6
1
Centre d’Etudes Biologiques de Chiz
e, UMR 7372, CNRS and Universit
e de La Rochelle, 79360 Beauvoir sur Niort,
France
2
Tour du Valat, Institut de recherche pour la conservation des zones humides M
editerran
eennes, 13200 Le Sambuc
Arles, France
3
Muséum National d'Histoire Naturelle, Centre d’Ecologie et des Sciences de la Conservation-CESCO - UMR 7204
MNHN-CNRS-UPMC, 43 rue Buffon, 75005 Paris, France
4
GeolinkX, P-A C^
ote Rousse, 180 rue du Genevois, 73000 Chamb
ery, France
5
INRA, Biostatistics & Spatial Processes (BioSP), Domaine Saint-Paul, Site Agroparc, 84914 Avignon, France
6
LTSER ‘Zone Atelier Plaine & Val de S
evre’, CNRS, 79360 Beauvoir sur Niort, France
Agricultural intensification over the past decades has led to a generalized decline in farm-
land biodiversity. Farmland birds are particularly exposed to rapid changes in habitat and
reduced food resources or availability. Understanding how farmland specialists can be
preserved and their populations enhanced are major challenges for this century. Based
on a long-term (19-year) study of a Eurasian Stone-curlew Burhinus oedicnemus popula-
tion, we estimated the demographic parameters, including clutch size, egg volume,
hatching success, survival rate and apparent population size. Demographic rates found
for this French population were, on average, comparable to those found elsewhere in
Europe. However, all demographic parameters showed negative trends, including a dra-
matic decline in the local population (26% decline over 14 years) and a 10% decline in
adult survival rate over 11 years. Such a long-term decline, despite on-going conservation
efforts, calls into question the overall sustainability of arable Stone-curlew populations.
We infer some of the possible causes of this decline, in particular food shortage, and dis-
cuss how this pattern could be reversed through conservation measures applicable at
very large spatial scales.
Keywords: breeding, Burhinus oedicnemus, demographic rate, farmland birds, population
dynamics, population monitoring, protection status.
Agricultural expansion over the last 10 000 years
has created a complex mosaic of landscapes which
replaced primeval forests and steppe habitats
(Kaplan et al. 2009). Extensive farming allowed
the colonization of these new habitats by numer-
ous species of birds, usually of steppe origin
(O’Connor & Shrubb 1986). However, over the
past century, intensive agriculture has replaced tra-
ditional farming, a trend that has been accelerated
by the Common Agricultural Policy (CAP) in
Western Europe since 1962. CAP-induced changes
in agricultural practices are a major cause of farm-
land biodiversity loss, especially birds (Krebs et al.
1999, Donald et al. 2001). Farmland specialist
birds have been extensively studied to understand
the multifactorial causes of decline linked to inten-
sive farming practices (Aebischer & Ewald 2012,
Kentie et al. 2013, Chiron et al. 2014, Barr
eet al.
2018). However, the breeding ecology of several
farmland birds, including some threatened species,
and the detailed mechanisms by which they are
*Corresponding author.
Email: elie.gaget@gmail.com
© 2018 British Ornithologists’Union
Ibis (2018) doi: 10.1111/ibi.12646
affected by intensive farming practices, still remain
to be elucidated in many cases (Fuller et al. 1995,
Chamberlain et al. 2000, Heldbjerg et al. 2017,
Stanton et al. 2018).
The Eurasian Stone-curlew Burhinus oedicnemus
(Charadriiformes, Burhinidae; hereafter Stone-
curlew) is a steppic Palaearctic bird occurring in
European farmlands and pseudo-steppes (Vaughan
& Vaughan-Jennings 2005). The species suffered a
rapid and important population decline over the
second half of the last century (Cramp & Simmons
1983). However, despite the scarcity and impreci-
sion of national trend data, its European conserva-
tion status has remained in the ‘least concern’
category, with an estimated 53 400–88 200 pairs
in the EU (BirdLife International 2017). Indeed,
the population trend is unknown for 46% of Euro-
pean countries, shows positive trends for only
14%, is stable or fluctuating for 21% and is nega-
tive for 18% (BirdLife International, 2017). Apart
from countries where the species is highly local-
ized (e.g. in the UK), trends should probably be
best considered tentative. A decrease in geographi-
cal range and breeding population was reported in
France over the second half of the 20th century
(Yeatman-Berthelot & Jarry 1994), with an esti-
mated breeding population of 5000–9000 pairs
between 1980 and 1993 (Malvaud 1996), with
most recent estimates for the French population
size of c. 19 000–28 000 breeding pairs (Issa &
Muller 2015). In the UK, the situation in the
1980s was almost desperate, but over the last
three decades, owing to a major conservation
effort of the RSPB (Evans & Green 2007), the
population reached c. 400 breeding pairs (Eaton
et al. 2011).
Stone-curlew population monitoring data are
scarce because there are very few long-term field
studies that could provide accurate trends, partly
because of the elusive behaviour, shyness and
excellent camouflage of the species. In addition,
Stone-curlew breeding habitat choice is surpris-
ingly flexible: any kind of habitat with drained
soils, low vegetation height and density, and stones
on the ground to optimize anti-predation strategies
for this cryptic species, seems to fulfil its habitat
requirements (Green et al. 2000). Breeding habitat
includes heathlands, semi-natural grasslands,
pseudo-steppes, gravel riverbeds, vineyards, orch-
ards, spring-sown crops and brownfields (Vaughan
& Vaughan-Jennings 2005). Conservation success
in the UK relied to a large extent on detailed
breeding biology and habitat selection studies
(Gibbons et al. 1996) which helped shape Agri-
Environmental Scheme (AES) implementation
(Grice et al. 2007). The latter mainly consisted of
nesting plots in an uncultivated area within spring-
sown crops of 1–2 ha, away from field boundaries
and near pastureland (Evans & Green 2007).
However, implementing this AES elsewhere in
Europe requires extended knowledge of the breed-
ing biology of the species, either in arable crops or
in more natural steppic or pseudo-steppe habitats.
France, with 21% of the European population,
represents the second largest European breeding
population after Spain (BirdLife International
2017). In France, farmland landscapes are the
major breeding habitat, with over 60% of breeding
pairs being located in arable crops of the central-
western region (Malvaud 1996, Issa & Muller
2015). In such habitat, however, the species is
threatened by nest destruction though agricultural
work (Berg et al. 2002, Whittingham & Evans
2004), and chick survival as well as adult fitness
are potentially threatened by a decrease in food
resources, which is known adversely to affect farm-
land specialists (Donald et al. 2001). The aim of
this study was to assess the status and trends (over
19 years) of demographic parameters of a Stone-
curlew population breeding in an intensive farm-
land landscape and benefiting indirectly from agri-
environmental conservation measures.
METHODS
Study area and conservation measures
The Long Term Social-Ecological Research site
(LTSER) ‘Zone Atelier Plaine & Val de S
evre’
(http://www.za.plainevalsevre.cnrs.fr/, Bretagnolle
et al. 2018) is located within an intensively man-
aged farmland area in the Poitou-Charentes
Region, Deux-S
evres district, central-western
France (Fig. 1). The site covers 450 km²of farm-
land, where crops are dominated by winter annuals
(cereals c. 40% and rapeseed c. 15% of the arable
surface), followed by spring crops (sunflower 15%
and maize 10%) and perennial covers (10%; Bre-
tagnolle et al. 2018). The plain lies upon a Jurassic
sedimentary basin, with well-drained and poor soil,
typical of a rendzina (INRA 1998).
Half of the LTSER was designated as a Special
Protected Area (SPA Natura2000, FR5412007,
207.6 km²) in 2004 due to the presence of 17
© 2018 British Ornithologists’Union
2E. Gaget et al.
species listed in Annex I of the Birds Directive.
Some AES have been implemented on the LTSER
since 1999, but since 2004, AES have been imple-
mented more strongly within the framework of
the CAP (Bretagnolle et al. 2011, Berthet et al.
2012). AES have been mainly targeted toward the
conservation of Little Bustards Tetrax tetrax and
consist of increasing grassland cover and fodder
crops, decreasing mowing frequency in alfalfa and
permanent grasslands from May to August in order
to limit nest destruction and the killing of incubat-
ing females, and banning pesticides to increase
food resources for the chicks (Bretagnolle et al.
2011). Up to 10 000 ha of contracts have been
established (Bretagnolle et al. 2011, Caro et al.
2016). These measures have increased the overall
amount of preserved nesting habitat as well as
food resources for many farmlands birds (Bretag-
nolle et al. 2011, Brodier et al. 2014). Stone-cur-
lews breeding on the LTSER have thus probably
benefited from habitats that were on average of
higher quality regarding food resources. In addi-
tion, since the beginning of monitoring in 1998,
the species has benefited from an awareness pro-
gramme aimed at farmers on a sub-site of 4300 ha
(the one where all breeding parameters were col-
lected for this study, see VA sub-site below): nests
found following intensive searches were reported
to farmers, and nest locations were marked in the
field, to avoid destruction during agricultural work.
Overall, measures to improve food availability as
well as nest protection were expected to maintain,
if not increase, reproductive investment (clutch
size, egg volume and hatching rate) and hence
population size on the VA sub-site.
Stone-curlew breeding biology
Breeders were monitored in a delimited sub-site
of the LTSER of c. 4300 ha (hereafter, VA), over
19 consecutive years (1998–2016). Each year,
from March to June, all fields with favourable
vegetation height (<15 cm) were monitored on a
weekly basis (approximately 200–400 fields cover-
ing 500–1200 ha). Nests were located from dis-
tant vantage points by using a telescope
(20 960 mm) and subsequently visited to deter-
mine breeding stage precisely. The first visit usu-
ally occurred before hatching, when egg biometric
measurements allowed the determination of laying
and hatching dates with the use of a calibration
density curve (Hoyt 1979, V. Bretagnolle unpubl.
data, Fig. S1). Egg weight, length and width were
recorded (precision of 0.1 g and 0.1 mm, respec-
tively; Table S1). Egg density (mass/volume),
which decreases during incubation (Green & Grif-
fiths 1994), was used to estimate egg laying date
at a precision of 1.52 days (V. Bretagnolle
unpubl. data, Fig. S1). Pairs were then re-checked
at least once a week to ensure they were still pre-
sent and incubating. The nests were re-visited if
pairs were not observed for two consecutive days,
or around the hatching date, to determine the
fate of the clutch (hatching, destruction by agri-
cultural work, desertion/predation). Because it
was not possible to determine with certainty
whether an empty nest had been deserted before
eggs were removed by a predator, we used a sin-
gle category ‘desertion/predation’. If at the first
nest visit, eggs had already hatched and chicks
were still close to the nest, the laying date was
retrospectively calculated with reference to the
incubation period of 26 days (Vaughan &
Vaughan-Jennings 2005) and the chicks’age (esti-
mated with a precision of 2.6 days with the use
of a wing measure calibration curve, V. Bretag-
nolle unpubl. data; Fig. S2).
Figure 1. Map of the study area, the Long-term Social-Ecolo-
gical Research Site (LTSER) ‘Zone Atelier Plaine & Val de
S
evre’. The grey polygons correspond to the four sub-sites
used for the survey of the Stone-curlew population, of which
‘VA’in light grey corresponds to the monitoring sub-site. The
dotted lines delimit the Special Protected Area (SPA Natu-
ra2000, FR5412007).
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 3
For each nest, laying date was thus obtained
(for two-egg clutches, the mean laying date) and
expressed in Julian calendar days, starting from 1
March. Clutches from July to early September
(<3% of recorded breeding events) were discarded
from the analyses because observation pressure
during those months varied over time. Stone-cur-
lews are able to lay replacement clutches after nest
failure, as well as true second clutches (Vaughan
& Vaughan-Jennings 2005). To estimate the num-
ber of breeding attempts per pair, we used a mix-
ture distribution method (log-normal, Bealey et al.
1999) using the observed distribution of the laying
dates (R package ‘mixdist’, Macdonald & Du
2012). We parameterized the model with the lay-
ing date of nests with ringed breeding birds
(n=130) and then fitted it on the complete data-
set (n=513). This analysis is a combination of the
Newton-type method and the estimate mean algo-
rithm (O’Neil 1971). The unconstrained model
finds a set of overlapping distributions of laying
dates (we used three log-normal distributions,
accounting for the possibility of three successive
clutches for a given pair) that provides the best fit
to grouped data. The quality of the model (com-
parison between observed data and estimated dis-
tributions) was tested with a Chi-square goodness-
of-fit test (Macdonald & Du 2012).
Temporal trends in breeding parameters were
then evaluated successively in different models.
Trends in laying date, clutch size and egg volume
were investigated using a generalized linear model
(GLM) with either a Gaussian error distribution
(laying date and egg volume) or a binomial error
distribution (clutch size, modelling the probability
of a one- or two-egg clutch). For the laying date,
we tested the temporal trend over years on two
subsets: over the breeding season (from March to
June) and over the first half of the first breeding
attempts, which does not include replacement
clutches (from March to 27 April). For egg vol-
ume and clutch size, we considered the effect of
laying date and its interaction with year (Christians
2002). Nest fate was investigated using the May-
field model based on a maximum likelihood
approach. Compared with the initial Mayfield
model (Mayfield 1975), no assumption about
when the failure occurs is required, and covariates
can be easily incorporated (Rotella 2014). We
used a multi-state model to include directly the
two identified causes of failure, i.e. ‘destroyed by
agricultural work’or ‘deserted/predated’(Darrah
et al. 2018). We tested whether nest survival and
causes of failure changed according to year, laying
date and clutch size considering linear relation-
ships. Additionally, we tested the effect of the
interaction between laying date and year. Nest sur-
vival from laying until hatching was calculated by
raising the daily survival rate to the power of 26
(i.e. the incubation period) and the corresponding
variance was estimated by the delta method (Pow-
ell 2007). Nest fate model building and parameter
estimates were obtained using E-SURGE v.1.8.5
(Choquet et al. 2009a). For all these models inves-
tigating the breeding parameters and their tempo-
ral trend, we used a model selection inference
with corrected Akaike information criterion
(AICc). The ability of two models to describe the
data was assumed to be identical if the difference
in their AICc was <2. However, in particular
cases where models within the two units of the
best model have only one more parameter, the lar-
ger model is not necessarily supported or competi-
tive. A closer examination considering the
deviance is required to see whether the fit is really
improved, or whether the model is ‘close’in terms
of AICc because it adds only one parameter (Burn-
ham & Anderson 2002). In the latter case, we
selected the most parsimonious model (i.e. that
with the lowest number of parameters).
Trend in apparent population size
To evaluate the trend in population size, two dif-
ferent methods were used. First, counts were per-
formed every year on four different sub-sites from
2003 to 2016, totalling 16 000 ha including the
monitoring VA sub-site (Fig. 1). All ploughed
fields (an area of c. 3000 ha), i.e. those sown with
sunflower or maize (vegetation height under
15 cm), were systematically inspected for 1–5 min
according to field size and topography. Observa-
tions were carried out at the beginning of May
over 8–15 days (the precise dates varied from year
to year according to spring crop growth). As the
LTSER is very varied in topography and cropping
systems, there were substantial differences
between the four sub-sites. Observations were
always performed in good sighting conditions (no
heavy rainfall or heat haze), usually at 07:00–
11:00 h and 16:00–20:00 h. As detection proba-
bility was not accounted for, we measured appar-
ent population size rather than true population
size. We used a GLM with a Poisson error-
© 2018 British Ornithologists’Union
4E. Gaget et al.
distribution (and log link) with a hypothesis testing
approach (i.e. based on P-value with a=0.05) to
test for the temporal trend in abundance, number
of pairs detected (simply defined as two birds seen
together in the same field) and number of fields
occupied by one or more birds. Explanatory vari-
ables included: sub-site identity (a factor with four
levels), year (as a continuous variable) as well as
their interaction. In addition, the surface of the
surveyed area and the number of fields surveyed,
which varied between years and sub-sites, were
entered as offsets after log-transformation. Because
75% of the observers were involved for only
1 year, and only 5% for more than 3 years, we did
not include observer identity in a mixed effect
modelling framework. We investigated whether
residuals displayed spatial autocorrelation thanks
to a spatial variogram (R package ‘spatial’, Ven-
ables & Ripley 2002). The exponential growth rate
of the abundance was extracted from the year
term.
The second monitoring survey relied on the
breeding biology monitoring scheme (see above)
and concerned only the VA sub-site, where a thor-
ough nest search was carried out every year from
March to June. The long-term trend in number of
nests was tested using a generalized additive model
with a smoothed term on the year (GAM, Gaus-
sian error distribution) and a hypothesis testing
approach (based on P-value with a=0.05). How-
ever, as the monitored period differed slightly
between years, we tabulated the number of pairs
for the extended period (15 March–30 June) and
for reduced, better standardized periods: 15
March–30 May and 1 April–10 May.
Survival rates
Stone-curlews were captured in the monitored
breeding sub-site VA from 2005 between March
and September. Birds were ringed with a metal
ring (National Museum of Natural History,
MNHN, Paris, France) and a combination of two
or four colour rings (http://cr-birding.org/node/
89). Chicks were ringed only if older than 10 days.
Between 2005 and 2015, 93 adults and 68 chicks/
fledglings were ringed, resulting in a total of 254
re-sightings. Adult bodyweight (g) and wing length
(mm) were measured for captures (n=57) and
recaptures (n=6). Body condition was estimated
using the scale mass index (SMI) that explicitly
accounts for the allometrical relations (Peig &
Green 2009). Body mass was standardized for a
given size using the following equation:
c
Mi¼MiL0
Li
bsma
Where c
Miis the predicted body mass for individual
iwhen the body measure is standardized to L
0
,an
arbitrary value of L.M
i
and L
i
are the body mass
and the body measurement of individual I, respec-
tively; b
sma
is the scaling exponent estimated.
Survival estimates were obtained by capture–re-
capture analysis using a Cormack–Jolly–Seber
model. Parameters directly estimated by the model
were /, the apparent survival probability, and p, the
re-sighting probability. To avoid over-parameteriza-
tion, we used a two-step model selection procedure.
First, we selected the best model structure based on
a full general model with an AICc-based model
selection. Secondly, we assessed the presence of a
trend over the study period on juvenile and adult
survival, and tested the possible effect of body con-
dition at capture year ton adult survival in year
t+1. In our general model, survival probability was
age- and sex-dependent. For the effect of age, we
distinguished two classes, juvenile (first year) and
adult (>1 year; Green et al. 1997). Re-sighting
probability was time- and sex-specific because
brooding is mainly performed during the day by the
female (preventing rings from being read), and the
male is predominantly in the ‘spotter’position
(V. Bretagnolle pers. obs.). We considered only an
additive effect of time for re-sighting probability
because an interactive effect with sex leads to an
over-parameterized model. Thus our general model
was /
juv.sex ad.sex
p
t+sex
where juvenile is denoted by
‘juv’, adult by ‘ad’, additive effect by ‘+’and interac-
tive effect by ‘.’. To assess the effect of body condi-
tion on survival, we included in the best model the
logistic regression: logit(/)=b0+b19x
i
, where
/is the survival probability the year following the
first capture, b0 is an intercept parameter, b1isa
slope parameter, and x
i
is the body condition of
individual iat first capture time. Model building,
model selection (AICc, Burnham & Anderson
2002) and parameter estimates were obtained using
E-SURGE (v.1.8.5, Choquet et al. 2009a). The
model selection method was identical to that pre-
sented above (see section on breeding biology). Fol-
lowing Grosbois et al. (2008), we used a likelihood
ratio test (LRT, hypothesis testing approach,
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 5
a=0.05) to estimate the significance of a trend in
survival, because residual survival variation is null
after integrating. We performed goodness-of-fit
(GOF) tests using the program U-CARE (v.2.3.2,
Choquet et al. 2009b). Finally, the temporal trend
in body condition index was tested under a
hypothesis testing approach (a=0.05) with a lin-
ear mixed effect model (LMM), with year and
date of capture as fixed effects and individual as a
random effect.
All statistical analyses were run in R 3.2.0 (R
Development Core Team 2015). For GLM and
LMM, residuals of the models were checked using
graphic methods to verify the assumptions of nor-
mality, non-overdispersion and homoscedasticity.
Means are presented sd unless stated otherwise.
RESULTS
Breeding biology
Over the 19 survey years, 566 nests were found,
of which 513 provided an estimated laying date.
Nests with at least one ringed bird (n=130)
allowed us to confirm the existence of true second
clutches after a successful attempt (i.e. double
brooding, n=5) and even that of pairs having
three successive breeding attempts (n=2). After
successful fledging (at the age of 50 days, Green
et al. 1997) or a breeding failure, a new reproduc-
tive attempt was started on average
13.5 4.2 days later (range 10–20 days, n=5).
The first peak of laying dates was around 18 April
(Fig. 2, median of the log-normal distribution
around 27 April 17 days), with the earliest
clutch being laid on 15 March. The second peak
of laying dates was around 22 May (Fig. 2, median
of the log-normal distribution around 25
May 11 days). A few late clutches were laid by
the end of June (Fig. 2). Some laying occurred up
to mid-September, although these were not
included in the analysis for protocol consistency
(<3% of nests). The mixture distribution model
provided an estimate of 1.17 0.11 breeding
attempts per pair (v²
5
=5.66, P=0.34). Most
clutches (85.0%, n=533) were two-egg clutches
(mean 1.85 0.36 eggs), although some one-egg
clutches may have been two-egg clutches subject
to accidental loss or partial predation. On average,
two- and one-egg clutches were visited respec-
tively 9.9 6.9 and 13.0 7.7 days after the lay-
ing date. In addition, the first visit occurred in the
first 3 days after laying for 24% and 14% of the
two- and one-egg clutches, respectively.
On average, raw hatching success was 53 15%,
which was reduced to 32 3% after correction
using the Mayfield method (n=441). Desertion/
predation accounted for 85 3% of clutch failures,
and direct destruction caused by mechanical agricul-
tural work caused 15 3% of failures. Once nests
were discovered, however, they were marked and
farmers were immediately informed and asked to
avoid them during farm work. The proportion of
nests lost due to agricultural activity is expected to
be higher during the period between egg-laying and
nest detection, and therefore nest destruction from
sowing or hoeing was probably underestimated.
There was no strong evidence that clutch size had an
effect on nest survival or on the cause of failure (i.e.
each Akaike information criterion size had
DAICc >2, Tables 1 and S2).
Long-term and seasonal trends in
breeding parameters
Model selections provided support for a temporal
trend for all breeding parameters, except laying
date (Table 1). We found that nest survival and
Figure 2. Distribution of laying dates (histogram, n=513,
1998–2016), of the fitted log-normal distributions (dotted lines,
see Methods) and of the total fitted laying date (thick line). The
x-axis represents the laying date in Julian days since 1 March
(31 =1 April, 61 =1 May, 92 =1 June). The laying peaks are
around 18 April and 22 May, and the medians (triangles on
the x-axis) are around 27 April and 25 May.
© 2018 British Ornithologists’Union
6E. Gaget et al.
egg volume had declined over the study duration
by 80% and 2%, respectively (Fig. 3). Decrease in
nest survival over the years was the consequence
of increasing desertion/predation rate, which var-
ied from 30 9% at the beginning of the study
period (1998) to 80 4% during the last years
(2014-2016, Fig. S3). Over the same period,
destruction rate was relatively stable at around
11 4% (Fig. S3). Results suggested an increase
in clutch size over time, but evidence for this
trend was weak, as the constant model was also
in the best model set (DAICc =0.8). Within
years, we found clear support for a seasonal trend
in all breeding parameters, except clutch size
(Tables 1 and S2). Nest survival and egg volume
decreased over the breeding season (Table 1).
The cause of failure also changed, with the high-
est proportion of nests lost due to desertion/pre-
dation for late clutches (Table 1, Fig. S4). Finally,
the results provided moderate support for a posi-
tive interaction between laying date and year on
nest survival and egg volume (DAICc =2 and 1,
respectively), suggesting that the negative seasonal
trend previously described has been attenuated
over the study duration (Table 1). An interaction
between laying date and year was not supported
as the cause of failure and clutch size (Tables 1
and S2).
Trends in apparent population size
Using data from the four sub-sites, we found that
abundance decreased significantly over the
14 years (Fig. 3, GLM, b=0.03, z=3.1,
P=0.002) with an exponential growth rate of
r=0.979, 95% CI: 0.958–0.989. The same results
were verified for the number of pairs detected
(GLM, b=0.03, z=2.0, P=0.05) and for
the number of occupied fields (GLM, b=0.03,
z=2.9, P=0.004). The sub-site effect was sig-
nificant as well as its interaction with year for the
three investigated variables (total abundance, num-
ber of pairs and number of occupied fields,
P<0.0001). In the VA sub-site, the values were
significantly higher and trends were more negative
(P<0.0001). The PR sub-site, outside of the
SPA, was not significantly different from the FO
and SB sub-sites in mean or interaction effects
(P>0.4). No significant linear or polynomial
trends were detected for the number of nests
Table 1. Results from model selection testing for a linear effect of time (period 1998–2016) on laying date (LD), nest survival, nest
desertion/predation probability given failure, clutch size and egg volume.
Model k DEV AICc DAICc Slope 1 Slope 2 Interaction
Laying date (March–June)
Constant model 1 4492.6 4496.7 0.0
Year 2 4491.8 4497.8 1.1 0.21 0.14
Laying date (March–27 April)
Constant model 1 1702.2 1706.2 0.0
Year 2 1701.2 1707.2 1.0 0.10 0.10
Nest survival
Year +LD +Year:LD 10 1245.0 1265.0 0.0 0.95 0.27 0.42 0.16 0.53 0.27
Year +LD 9 1249.0 1267.0 2.0 0.44 0.08 0.15 0.08
Nest desertion/predation probability given failure
Year +LD 10 1245.0 1265.0 0.0 0.79 0.25 0.85 0.25
Year +LD +Year:LD 11 1244.7 1266.8 1.8 0.39 0.79 0.63 0.47 0.49 0.92
Clutch size
Year 2 399.3 403.4 0.0 0.22 0.13
Constant model 1 402.2 404.2 0.8
Year +LD 3 398.9 404.9 1.5 0.22 0.13 0.09 0.13
Egg volume
Year +LD +Year:LD 4 4052.2 4062.3 0.0 0.23 0.10 0.57 0.10 0.19 0.09
Year +LD 3 4055.2 4063.2 1.0 0.22 0.10 0.57 0.10
Tested variables include the laying date (LD) for all breeding parameters, excluding the laying date, and clutch size for nest survival
and nest desertion/predation probability given failure. For each model, results include the number of parameters (k), deviance (DEV),
AIC value corrected for small sample size (AICc), difference between current model and the best model within each sub-set of model
(DAICc) and the estimated slope and intercept se of the highest ranked model. ‘+’indicates an additive effect and ‘:’an interaction.
All covariates were standardized. Only the highest ranked models (DAICc ≤2) are shown. Details of other models are given in
Table S2.
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 7
found at the VA sub-site, irrespective of the sur-
vey period retained (GAM, P>0.05; Fig. S5).
Survival rates
Goodness of fit tests (v
2
=52.7, P=0.60) pro-
vided no indication of lack of fit. The first step of
model selection suggested that resighting probabil-
ity was time- and sex-specific (Table 2). As
expected, males had an average resighting proba-
bility higher than that for females (males 0.76,
95% CI: 0.62–0.86; females 0.53, 95% CI: 0.38–
0.67). We did not find any evidence for sex-speci-
fic survival rates, but there was strong support for
different apparent survival rates between juveniles
and adults (Table 2, M5 vs. M8 DAICc =14.92).
Juvenile survival was 0.55 (95% CI: 0.41–0.69)
and adult survival was 0.88 (95% CI: 0.83–0.91).
For the second step of the model selection, LRT
supported a linear trend in adult survival over the
study period (F
cst/trend/t
=8.11, P=0.004;
Table S3), with an average decrease in apparent
survival of 2.3% per year (Fig. 3). Finally, adult
body condition index (average mass =490.0
38.6 g) decreased, but not significantly, over the
study period (LMM, b=2.00, t
1,89
=1.67,
P=0.23). We found a positive relationship
between body condition and female adult survival
(Table S3).
DISCUSSION
Our study provides detailed information on the
breeding biology and population trends of the
Stone-curlew using one of the longest time-series
available, and the only available one for France. In
addition, the study was located in one of the
strongholds of the species, the Poitou-Charentes
region, which harbours c.13–21% of the French
population (Issa & Muller 2015). Our results indi-
cate a long-term decline in this population. Such a
decline, despite on-going conservation efforts, calls
Figure 3. Trends in (a) population size (four sub-sites: cross for VA, triangle for PR, circle for FO and square for SB), (b) adult
apparent survival rate se, (c) egg volume se and (d) nest survival se. The predicted values were extracted from the corre-
sponding GLMs (see Methods) and shown with their 95% CI. All trends are significant.
© 2018 British Ornithologists’Union
8E. Gaget et al.
into question the overall sustainability of arable
Stone-curlew populations.
Breeding success and survival rate in an
intensive farmland landscape
Within a European context, the observed nest sur-
vival rate and survival of individuals in this study
are comparable to those obtained for the UK,
Spain and Italy (Table S4). Nest destruction due
to agricultural work was responsible for 11% of
nest failures in our study, mainly occurring at a
very early stage of incubation, i.e. before signalling
the presence of the nest to the farmers, and which
is underestimated. The rate of nest destruction
without protection was estimated at 33% in 2001
and nearly 50% in 2012 (V. Bretagnolle unpubl.
data), thus constituting the major threat encoun-
tered by the species in such farmland habitat.
Additionally, predation has been reported as the
main cause of nest failure (Solis & Lope 1995,
Bealey et al. 1999) and probably accounts for most
of the desertion/predation events reported in this
study, even if the effect of crop growth was not
estimated. Some one-egg clutches could also have
resulted from partial egg predation before the first
nest visit. The increase in desertion/predation over
time could result either from a reduction in nest
protection, given that parents in weak body condi-
tion cannot ensure proper parental care such as
nest defence or nest attendance after predator
encounters (Winkler 1992), or by an increase in
the predator populations.
The long-term decline of the
Stone-curlew in intensive farmlands
All investigated demographic parameters displayed
negative trends over time. Although apparent
rather than true survival rate was estimated, which
may not exclude permanent emigration from the
study area (an unlikely scenario given the species
is known to be highly philopatric; Green 1990), a
decrease in adult survival is of concern for popula-
tion stability, as population growth rate is highly
sensitive to adult mortality in long-lived species
(Sæther & Bakke 2000). Although this study took
place at a relatively small spatial scale (c.
4300 ha), which may limit the generality of the
conclusions, this population benefitted from AES
dedicated to the preservation of trophic resources
for farmland birds (Bretagnolle et al. 2011) and
from active nest protection from agricultural work.
Consequently, the decrease in breeding success
and survival which resulted in a rapid population
decline (26% in 14 years) occurred in what could
be described as the best current possible conditions
for the species in intensive French agricultural
landscapes.
Which factors, affecting both survival and
breeding process, may have caused the population
decline? As suggested, nest destruction during sow-
ing or mechanical weeding is a well-known major
factor, but with limited impact in our case thanks
to the nest awareness programme. In addition, we
suggest that food limitation may play an important
but often overlooked role. Of particular interest in
this respect is the decrease in egg volume (2% in
19 years), despite AES implementation since the
first years of monitoring, enhancing overall habitat
quality (Bretagnolle et al. 2011). Within a given
season, a decrease in egg volume is found in many
bird species, as in our population, due to early
breeders being of higher quality than late breeders
(Christians 2002, Verhulst & Nilsson 2008). How-
ever, food availability can also affect egg volume
(Robb et al. 2008). Agricultural intensification is
considered a key factor which adversely impacts
the diversity and abundance of insects (Donald
et al. 2001, Johnson 2007, Geiger et al. 2010).
Table 2. Survival (/) and resighting (p) modelling as a func-
tion of age and sex between 2005 and 2015.
No. Model k DEV AICc DAICc
Resighting
1p
sex+t
22 797.43 844.02 0
2p
sex
10 826.31 846.85 2.84
3p
t
21 802.57 846.93 2.91
4p
cst
9 831.31 849.76 5.74
Survival: sex effect
5/
juv_ad
20 798.07 840.21 0
6/
juv.sex_ad
21 797.76 842.11 1.90
7/
juv_ad.sex
21 797.80 842.16 1.95
1/
juv.sex_ad.sex
22 797.43 844.02 4.78
Survival: age effect
5/
juv_ad
20 798.07 840.21 0
8/
cst
19 815.20 855.13 14.92
Results of model selection include: number of mathematical
parameters (k), the deviance (DEV), AIC value corrected for
small sample size (AICc) and difference between the current
model and the best model within each sub-set of models
(DAICc). The final selected model is in bold type. For model
notation, ‘juv’indicates juvenile, ‘ad’indicates adult, ‘cst’indi-
cates a constant parameter, ‘+’indicates an additive effect and
‘.’indicates an interactive effect.
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 9
Recent studies have demonstrated a relationship
between widespread application of pesticides,
neonicotinoids in particular, with concomitant
declines in insect and plant communities, and
decreases in insectivorous or granivorous birds
(Mineau & Whiteside 2013, Hallmann et al. 2014,
Gilburn et al. 2015). The diet of the Stone-curlew
is based on earthworms and beetles (Amat 1986,
Green et al. 2000). Even though a detailed analysis
of food availability and diet may be lacking in our
study, it is perhaps relevant to note that Poecilus
cupreus, the most abundant carabid species in our
study site (Marrec et al. 2015), has shown an aver-
age 80% decline in 20 years (V. Bretagnolle un-
publ. data). A further mechanism which may be
involved in the decrease in adult survival is that of
carry-over effects in wintering areas (Harrison
et al. 2011). Preliminary data from GPS tracking
of our breeding population has indicated a fairly
high diversity of wintering sites (France, Portugal
and Morocco, V. Bretagnolle & Groupe Ornitholo-
gique des Deux-S
evres unpubl. data).
Implications for conservation
The French Stone-curlew population has been
claimed to have increased in the period 2001–
2011 (BirdLife 2017). We question this conclu-
sion, especially given the absence of standardized
and dedicated protocols to monitor Stone-curlews
in France, and the cryptic nature of the species.
We suggest these positive trends actually result
from an increase in survey quality, i.e. a better
knowledge of the species habitat, and better data
transfer from observers (Issa & Muller 2015).
Based on our results, we suggest that Stone-cur-
lews breeding in farmland habitats may be cur-
rently in decline. Indeed, many farmland birds,
especially the largest species, are currently highly
threatened. For some, a dedicated AES framework
has proved useful (e.g. Verhulst et al. 2007, Bre-
tagnolle et al. 2011), despite AES having been
much criticized in the early years of its implemen-
tation (Kleijn & Sutherland 2003). AES schemes
dedicated to Stone-curlews are non-existent in
France, whereas in the UK, such AES schemes
consist of fallow plots providing suitable breeding
and foraging areas (Natural England 2010). In our
study area, however, we expected the species to
have potentially benefited from AES dedicated to
the Little Bustard. Such practices may provide
suitable habitat for the Stone-curlew and improve
food availability (Bretagnolle et al. 2011, Caro
et al. 2016). However, despite the fact that up to
10 000 ha of AES were established within the
LTSER (43 000 ha), Stone-curlew demographic
parameters have been declining. AES at a field
scale may not be appropriate because this species
forages over large areas (Green et al. 2000). Nev-
ertheless, AES for Stone-curlews can work, as the
RSPB Stone-curlew programme has successfully
demonstrated (Evans & Green 2007). However,
the success of this latter programme required a
strong investment in fieldworkers, applied
research, networking and funds for some hundreds
of breeding pairs.
In our case, there are possible efficient conser-
vation measures that could be implemented at a
far larger spatial scale. Given the current restricted
knowledge of this species, there is a clear and
urgent need to evaluate accurately whether this
decline is general (at the nationwide scale) or
restricted to some specific agricultural areas.
Implementing long-term monitoring of demogra-
phy and breeding parameters in this and other
French populations is therefore needed to assess
the potentially widespread and generic decline of
the species, not only in arable farmland landscapes,
but also in all semi-natural or artificial habitats.
This would require an assessment of: (1) the spa-
tial distribution and population size at the country
scale; (2) local/regional population trends in sev-
eral habitats; (3) diet in and outside the breeding
season; (4) the effects of predation and human dis-
turbance; (5) exposure to pesticides; and (6)
migratory strategies.
The creation of safe habitats to reduce brood
destruction and promote food availability, based on
the UK experience (Thompson et al. 2004), should
be explored. If similar patterns of population
decline were to be confirmed in other parts of
France, such conservation plots should be estab-
lished over hundreds of thousands of hectares to be
efficient, given the very large breeding distribution
of Stone-curlew. It should be also adapted to a
range crops (e.g. maize, sunflower, grasslands, vine-
yard). This would be challenging, as it would
require either pro-active campaigns targeting farm-
ers to adopt voluntary practices, or a consistent
funding scheme to compensate for the potential
yield loss to farmers at very large scales (Evans &
Green 2007), in a context of budgetary restriction
in agricultural subsidies. Alternatively, we may tar-
get the species environment and habitat rather than
© 2018 British Ornithologists’Union
10 E. Gaget et al.
the species itself. For instance, to improve food
availability, a reduction or ban of inputs may be
targeted. Some AES, organic farming or the recent
complete ban of neonicotinoids in France may
help. An increase of perennial crops, such as grass-
lands, and the enhancement of more extensive
practices should be strongly promoted as they sup-
port higher prey resources (Bretagnolle et al. 2011,
Badenhausser & Cordeau 2012, Caro et al. 2016) .
We would like to thank C. Atti
e, V. Rocheteau, M.
Liaigre, R. Bonnet, A. Millon, A. Desternes as well as all
fieldworkers who have helped to collect these data over
the years and spent much time in the field, and L. Inch-
board for improving the English text. Partial funding
was allocated by DREAL Poitou-Charentes and Fonda-
tion LISEA to V.B. We also thank the reviewers for
their very helpful comments and suggestions.
REFERENCES
Aebischer, N.J. & Ewald, J.A. 2012. The grey partridge in
the UK: population status, research, policy and prospects.
Anim. Biodivers. Conserv. 35: 353–362.
Amat, J.A. 1986. Information on the diet of the Stone Curlew
Burhinus oedicnemus in Do~
nana, southern Spain. Bird
Study 33:71–73.
Badenhausser, I. & Cordeau, S. 2012. Sown grass strip –A
stable habitat for grasshoppers (Orthoptera: Acrididae) in
dynamic agricultural landscapes. Agric. Ecosyst. Environ.
159: 105–111.
Barr
e, K., Le Viol, I., Julliard, R. & Kerbiriou, C. 2018.
Weed control method drives conservation tillage efficiency
on farmland breeding birds. Agric. Ecosyst. Environ. 256:
74–81.
Bealey, C.E., Green, R.E., Robson, R., Taylor, C.R. &
Winspear, R. 1999. Factors affecting the numbers and
breeding success of Stone Curlews Burhinus oedicnemus at
Porton Down, Wiltshire. Bird Study 46: 145–156.
Berg,
A., Jonsson, M., Lindberg, T. & K€
allebrink, K.G.
2002. Population dynamics and reproduction of Northern
Lapwings Vanellus vanellus in a meadow restoration area in
central Sweden. Ibis 144: 131–140.
Berthet, E.T., Bretagnolle, V. & Segrestin, B. 2012.
Analyzing the design process of farming practices ensuring
Little Bustard conservation: lessons for collective landscape
management. J. Sustain. Agric. 36: 319–336.
BirdLife International 2017. European Birds of Conservation
Concern: Populations, Trends and National Responsibilities.
Cambridge: BirdLife International.
Bretagnolle, V., Villers, A., Denonfoux, L., Cornulier, T.,
Inchausti, P. & Badenhausser, I. 2011. Rapid recovery of
a depleted population of Little Bustards Tetrax tetrax
following provision of alfalfa through an agri-environment
scheme. Ibis 153:4–13.
Bretagnolle, V., Berthet, E., Gross, N., Gauffre, B.,
Plumejeaud, C., Houte, S., Badenhausser, I., Monceau,
K., Allier, F., Monestiez, P. & Gaba, S. 2018. Towards
sustainable and multifunctional agriculture in farmland
landscapes: lessons from the integrative approach of a
French LTSER platform. Sci. Total Environ. 627: 822–834.
Brodier, S., Augiron, S., Cornulier, T. & Bretagnolle, V.
2014. Local improvement of Skylark and Corn Bunting
population trends on intensive arable landscape: a case
study of the conservation tool Natura 2000. Anim. Conserv.
17: 204–216.
Burnham, K.P. & Anderson, D.R. 2002. Model Selection and
Multimodel Inference: a Practical Information-theoretic
Approach, 2nd edn. New York: Springer.
Caro, G., Marrec, R., Gauffre, B., Roncoroni, M., Augiron,
S. & Bretagnolle, V. 2016. Multi-scale effects of agri-
environment schemes on carabid beetles in intensive
farmland. Agric. Ecosyst. Environ. 229:48–56.
Chamberlain, D.E., Fuller, R.J., Bunce, R.G.H., Duckworth,
J.C. & Shrubb, M. 2000. Changes in the abundance of
farmland birds in relation to the timing of agricultural
intensification in England and Wales. J. Appl. Ecol. 37: 771–
788.
Chiron, F., Charg
e, R., Julliard, R., Jiguet, F. & Muratet, A.
2014. Pesticide doses, landscape structure and their relative
effects on farmland birds. Agric. Ecosyst. Environ. 185: 153–
160.
Choquet, R., Rouan, L. & Pradel, R. 2009a. Program E-
SURGE: a software application for fitting multievent models
series. Environ. Ecol. Stat. 3: 845–865.
Choquet, R., Lebreton, J.D., Gimenez, O., Reboulet, A.-M.
& Pradel, R. 2009b. U-CARE: utilities for performing
goodness of fit tests and manipulating CApture-REcapture
data. Ecography 32: 1071–1074.
Christians, J.K. 2002. Avian egg size: variation within species
and inflexibility within individuals. Biol. Rev. 77:1–26.
Cramp, S. & Simmons, K.E.L. 1983. The Birds of the
Western Palearctic Volume III. New York, NY: Oxford
University Press.
Darrah, A.J., Cohen, J.B. & Castelli, P.M. 2018. A Bayesian
multinomial logistic exposure model for estimating
probabilities of competing sources of nest failure. Ibis 160:
23–35.
Development Core Team, R. 2015. R: A Language and
Environment for Statistical Computing. Vienna: R
Foundation for Statistical Computing.
Donald, P.F., Green, R.E. & Heath, M.F. 2001. Agricultural
intensification and the collapse of Europe’s farmland bird
populations. Proc. R. Soc. Lond. B 268:25–29.
Eaton, M.A., Balmer, D.E., Cuthbert, R., Grice, P.V., Hall, J.,
Hearn, R.D., Holt, C.A., Musgrove, A.J., Noble, D.G.,
Parsons, M., Risely, K., Stroud, D.A. & Wotton, S. 2011.
The State of the UK’s birds 2011. Sandy: RSPB, BTO,
WWT, CCW, JNCC, NE, NIEA & SNH.
Evans, A.D. & Green, R.E. 2007. An example of a two-tiered
agri-environment scheme designed to deliver effectively the
ecological requirements of both localized and widespread
bird species in England. J. Ornithol. 148: 279–286.
Fuller, J., Gregory, R.D., Gibbons, D.W., Marchant, J.H.,
Wilson, I.D., Baillie, S.R. & Carter, N. 1995. Population
declines and range contractions among lowland farmland
birds in Britain. Cons. Biol. 9: 1425–1441.
Geiger, F., Bengtsson, J., Berendse, F., Weisser, W.W.,
Emmerson, M., Morales, M.B., Ceryngier, P., Liira, J.,
Tscharntke, T., Winqvist, C., Eggers, S., Bommarco, R.,
P€
art, T., Bretagnolle, V., Plantagenest, M., Clement, L.W.,
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 11
Dennis, C., Palmer, C., O~
nate, J.J., Guerrero, I., Hawro, V.,
Aavik, T., Thies, C., Flohre, A., H€
anke, S., Fischer, C.,
Goedhart, P.W. & Inchausti, P. 2010. Persistent negative
effects of pesticides on biodiversity and biological control
potential on European farmland. Basic App. Ecol. 11:97–105.
Gibbons, D.W., Avery, M., Baillie, S., Gregory, R., Kirby, J.,
Porter, R., Tucker, G. & Williams, G. 1996. Bird Species of
Conservation Concern in the United Kingdom, Channel Isles
and Isle of Man: Revising the Red Data List. Sandy: RSPB
Conservation Review No. 10.
Gilburn, A.S., Bunnefeld, N., Wilson, J.M., Botham, M.S.,
Brereton, T.M., Fox, R. & Goulson, D. 2015. Are
neocotinoid insecticides driving declines of widespread
butterflies? PeerJ 3: e1402.
Green, R.E. 1990. Saving the stone curlew. Sanctuary Bull.
19:37–39.
Green, R.E. & Griffiths, G.H. 1994. Use of preferred nesting
habitat by Stone Curlews Burhinus oedicnemus in relation to
vegetation structure. J. Zool. 233: 457–471.
Green, R.E., Hodson, D.P. & Holness, P.R. 1997. Survival
and movements of Stone Curlews Burhinus oedicnemus
ringed in England. Ring. Migr. 18:24–34.
Green, R.E., Tyler, G.A. & Bowden, C.G.R. 2000. Habitat
selection, ranging behavior and diet of the stone curlew
(Burhinus oedicnemus) in southern England. J. Zool. 250:
161–183.
Grice, P., Radley, G., Smallshire, D. & Green, M. 2007.
Conserving England’s arable biodiversity through agri-
environment schemes and other environmental policies: a
brief history. Aspects Appl. Biol. 81:7–22.
Grosbois, V., Gimenez, O., Gaillard, J.-M., Pradel, R.,
Barbraud, C., Clobert, J., Møller, A.P. & Weimerskirch, H.
2008. Assessing the impact of climate variation on survival
in vertebrate populations. Biol. Rev. 83: 357–399.
Hallmann, C.A., Foppen, R.P.B., van Turnhout, C.A.M., de
Kroon, H. & Jongejans, E. 2014. Declines in insectivorous
birds are associated with high neonicotinoid concentrations.
Nature 511: 341–343.
Harrison, X.A., Blount, J.D., Inger, R., Norris, D.R. &
Bearhop, S. 2011. Carry-over effects as drivers of fitness
differences in animals. J. Anim. Ecol. 80:4–18.
Heldbjerg, H., Sunde, P. & Fox, A.D. 2017. Continuous
population declines for specialist farmland birds 1987-2014
in Denmark indicates no halt in biodiversity loss in
agricultural habitats. Bird Conserv. Int. 28: 278–292.
Hoyt, D.F. 1979. Practical methods of estimating volume and
fresh weight of bird eggs. Auk 96:73–77.
INRA 1998. La base de donn
ees g
eographique des sols de
France au 1/1 000 000. Paris: INRA, Service d’
etude des
sols et de la carte p
edologique de France.
Issa, N. & Muller, Y. 2015. Atlas des oiseaux de France
m
etropolitaine; nidification et pr
esence hivernale. Paris:
Delachaux & Niestl
e.
Johnson, M.D. 2007. Measuring habitat quality: a review.
Condo 109: 489–504.
Kaplan, O.J., Krumhardt, K.M. & Zimmermann, N. 2009.
The prehistoric and preindustrial deforestation of Europe.
Quat. Sci. Rev. 28: 3016–3034.
Kentie, R., Hooijmeijer, J.C., Trimbos, K.B., Groen, N.M. &
Piersma, T. 2013. Intensified agricultural use of grasslands
reduces growth and survival of precocial shorebird chicks. J.
Appl. Ecol. 50: 243–251.
Kleijn, D. & Sutherland, W. 2003. How effective are
European agri-environmental schemes in conserving and
promoting biodiversity? J. Appl. Ecol. 40: 947–963.
Krebs, J.R., Wilson, J.D., Bradbury, R.B. & Siriwardena,
G.M. 1999. The second silent spring? Nature 400: 611–612.
Macdonald, P. & Du, J. 2012. mixdist: finite mixture
distribution models. R package version 0.5-4. Available at:
http://cran.r⁃project. org/web/packa⁃ges/mixdist/index.html.
(accessed 16 May 2011) (08 February 2013)
Malvaud, F. 1996. L’Oedicn
eme criard en France, (Burhinus
oedicnemus). Caen: Groupe Ornithologique Normand.
Marrec, R., Badenhausser, I., Bretagnolle, V., B€
orger, L.,
Roncoroni, M., Guillon, N. & Gauffre, B. 2015. Crop
succession and habitat preferences drive the distribution
and abundance of carabid beetles in an agricultural
landscape. Agric. Ecosyst. Environ. 199: 282–289.
Mayfield, J. 1975. Suggestions for calculating nest success.
Wilson Bull. 87: 456–466.
Mineau, P. & Whiteside, M. 2013. Pesticide acute toxicity is a
better correlate of U.S. grassland bird declines than
agricultural intensification. PLoS ONE 8: e57457.
Natural England 2010. Entry Level Stewardship:
Environmental Stewardship Handbook, 3rd edn. Sheffield:
Natural England. February 2010.
O’Connor, R.J. & Shrubb, M. 1986. Farming and Birds.
Cambridge: Cambridge University Press.
O’Neill, R. 1971. Algorithm AS 47. Function minimization
using a simplex procedure. Appl. Stat. 20: 338–345.
Peig, J. & Green, A.J. 2009. New perspectives for estimating
body condition from mass/length data: the scaled mass
index as an alternative method. Oikos 118: 1883–1891.
Powell, L.A. 2007. Approximating variance of demographic
parameters using the delta method: a reference for avian
biologists. Condor 109: 949–954.
Robb, G.N., McDonald, R.A., Chamberlain, D.E. & Bearhop,
S. 2008. Food for thought: supplementary feeding as a
driver of ecological change in avian populations. Front. Ecol.
Environ. 6: 476–484.
Rotella, J.J. 2014. Nest survival models. In Cooch, E. &
White, G., eds. Program MARK: A gentle introduction.
Chapter 17: 20 p. Available at: http://www.phidot.org/softwa
re/mark/docs/book/ (accessed 5 May 2018).
Sæther, B.E. & Bakke, Ø. 2000. Avian life history variation
and contribution of demographic traits to the population
growth rate. Ecology 81: 642–653.
Solís, J.C. & Lope, F. 1995. Nest and egg crypsis in the
ground-nesting Stone Curlew Burhinus oedicnemus. J.
Avian Biol. 26: 135–138.
Stanton, R.L., Morrissey, C.A. & Clark, R.G. 2018. Analysis
of trends and agricultural drivers of farmland bird declines in
North America: a review. Agric. Ecosyst. Environ. 254: 244–
254.
Thompson, S., Hazel, A., Bailey, N., Bayliss, J. & Lee, J.T.
2004. Identifying potential breeding sites for the Stone
Curlew (Burhinus oedicnemus) in the UK. Agric. Ecosyst.
Environ. 12: 229–235.
Vaughan, R. & Vaughan-Jennings, N. 2005. The Stone
Curlew Burhinus oedicnemus. Falmouth: Isabelline Books.
Venables, W.N. & Ripley, B.D. 2002. Modern Applied
Statistics with S, 4th edn. New York, NY: Springer.
Verhulst, S. & Nilsson, J.-
A. 2008. The timing of birds’
breeding seasons: a review of experiments that manipulated
© 2018 British Ornithologists’Union
12 E. Gaget et al.
timing of breeding. Philos. Transact. R. Soc. Lond. 363:
399–410.
Verhulst, J., Kleijn, D. & Berendse, F. 2007. Direct and indirect
effects of the most widely implemented Dutch agri-environment
schemes on breeding waders. J. Appl. Ecol. 44:70–80.
Whittingham, M.J. & Evans, K.L. 2004. The effects of habitat
structure on predation risk of birds in agricultural
landscapes. Ibis 146: 210–220.
Winkler, D.W. 1992. Causes and consequences of variation in
parental defence behaviour by tree swallows. Condor 94:
502–520.
Yeatman-Berthelot, D. & Jarry, G. 1994. Nouvel Atlas des
Oiseaux nicheurs de France. 1985-1989. Paris: Soci
et
e
Ornithologique de France.
Received 29 May 2017;
revision accepted 24 June 2018.
Associate Editor: Niall Burton.
SUPPORTING INFORMATION
Additional supporting information may be found
online in the Supporting Information section at
the end of the article.
Table S1. Egg biometrics of the two-egg (smallest
and largest egg) and one-egg clutches (single egg).
Table S2. Details of the model selection on laying
date (LD), nest survival, nest desertion/predation
probability given failure, clutch size and egg volume.
Table S3. Testing for decreasing juvenile and adult
survival from 2005 to 2015 and body condition
effect on adult survival.
Table S4. Demographic parameters of Stone-cur-
lew populations in the European context.
Figure S1. Calibration density curve for eggs, to
determine the laying date with the egg biometric mea-
surements (data collected in captivity or directly in the
field, V. Bretagnolle, unpubl. data; Augiron 2007).
Figure S2. Calibration wing length curve for
chicks, to determine the hatching date (data collected
in captivity or directly in the field, V. Bretagnolle,
unpubl. data; Augiron 2007).
Figure S3. Desertion/predation and destruction
rate over the study duration.
Figure S4. Proportion of nest failure due to deser-
tion or predation as a function of laying date (in com-
parison to the nest failure due to destruction by
agricultural work).
Figure S5. Trendinnumberofnestsduringthe
three breeding period subsets in VA.
© 2018 British Ornithologists’Union
Stone-curlew decline in intensive farmlands 13
