Assessing the effects of trapping on pest bird species at the country level

Biological Conservation 158 (2013) 98–106

Contents lists available at SciVerse ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/biocon

Assessing the effects of trapping on pest bird species at the country level François Chiron ⇑, Romain Julliard Muséum national d’Histoire naturelle, UMR 7204 MNHN-CNRS-UPMC, 55 rue Buffon, 75005 Paris, France

a r t i c l e

i n f o

Article history: Received 27 February 2012 Received in revised form 24 July 2012 Accepted 1 August 2012 Available online 28 November 2012 Keywords: Pest control Corvids Breeding bird survey Extinction probability Age structure Robust design Detection probability France

a b s t r a c t Control to limit damage caused by undesirable organisms at the country level is a common management practice but its effects on the target populations are usually unknown. Monitoring consequences of control is however important to design and measure the efficacy of long-term management. Using data from the French Breeding Bird Survey and methodology that cope with detection bias, we studied the consequences of trapping on the age structure and spatial dynamics of the magpie (Pica pica), a bird considered as a pest species in France. Our results show that magpie occurrence in farmlands and semi-natural landscapes decreases with regional trapping intensity. Trapping increase the probability of populations becoming extinct locally, with less possibility of (re)colonising managed areas. Local extinction is likely the consequences of changes in the age structure of breeding populations which are composed of more immatures in intensively trapped areas. The effects of trapping are however mitigated in urban areas, which have become a refuge habitat for magpies. Trapping is a long established and very common practice in France. Although trapping has a successful impact on the magpie in countryside, it is recommended only if justified by conservation specific purposes. Non-lethal methods exist like reducing availability of human-related food resources, especially in urban environments. Monitoring the dynamics behind species occurrence is a useful approach to understand how control affects species distribution. This study illustrates the value of a national monitoring scheme in helping to understand trapping consequences. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Animals causing problems to human activities and conservation of other organisms are usually viewed as undesirable (Ormerod, 2002). In order to limit their impacts, populations of these species are managed (Côté and Sutherland, 1997). Because many management projects aim solely to eliminate these undesirable organisms (Conover, 2002), the contribution of science to this process is poor. Consequently, we know little about the efficiency of control activities and the impact on demography and distribution of target populations (Virgós and Travaini, 2005; Rushton et al., 2006; Zipkin et al., 2009). The effects of management activities on target species (e.g. pests and invasive species) are usually assessed at relatively small spatial scale units such as nature reserves or small game properties where control is a common management practice (Virgós and Travaini, 2005; Treves, 2009; but see Whitfield et al., 2007). At this scale, the effectiveness of control activities depends, for example, on the characteristics of capturing methods, the efforts expended and the skills of practitioners (Díaz-Ruiz et al., 2010), as well as

⇑ Corresponding author. Tel.: +33 1 40 79 58 53; fax: +33 1 40 79 38 35. E-mail address: [email protected] (F. Chiron). 0006-3207/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biocon.2012.08.001

on the characteristics of the target species (Villafuerte et al., 1998; Tryjanowski et al., 2009; Servanty et al., 2011). At this scale, the impacts of management on abundant and widespread species are usually short-lived because the control only covers a small extent of the species range (Harding et al., 2002; Beja et al., 2009). Animal populations can compensate for losses with the arrival of new individuals or changes in the demography (McDonald and Harris, 2002; Novaro et al., 2005). But culling can also create local population sinks if managed areas continue to attract individuals that are then systematically eliminated (Baker and Harris, 2006; Péron et al., 2012). In the long term, culling repeated in many different sites could have thus important indirect effects on neighbouring populations (Heydon and Reynolds, 2000). Control could impact targeted as well as non targeted populations and result in the modification of species range at large scale such as a country. This is likely to occur insofar as considerable efforts are made to control populations of some birds and mammals in Europe and elsewhere in the belief that this reduces their impact on game or threatened species (Villafuerte et al., 1998). It is often assumed that such activities should not cause any long-term decline in target populations unless the culling is enduring (Harding et al., 2002; McDonald and Harris, 2002). So far, there has been no assessment of large-scale control programs on the distribution and demography of target species, especially when they are undesirable.

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

Understanding how targeted populations respond to control on a large scale has probably been hampered by a lack of suitable data and appropriate methodology to describe relevant predictors. In recent years, there were more researches using large-scale monitoring schemes, such as national breeding bird surveys (BBSs) (Gregory et al., 2005; Newson et al., 2008). Although these surveys only yield a description of trends and distribution of populations, they can be of great value if population changes are combined with information on the predictors of these changes (e.g. changes in habitat use, fragmentation, climate; Gregory and Baillie, 1998; Julliard et al., 2004; Devictor et al., 2008). To our knowledge, no study has assessed the potential of monitoring to help evaluate the effectiveness of control activities on wildlife at a country scale. In France, a total of six bird species are listed as pest species (ROC, 2000) that include the Rook (Corvus frugilegus), the Carrion crow (Corvus corone), the Eurasian jay (Garrulus glandarius), the European starling (Sturnus vulgaris), the Woodpigeon (Columba palumbus) and the Black-billed magpie (Pica pica, hereafter the magpie). The usual and long-established response to limit impacts of these species is to kill the birds, usually by shooting or trapping. Yet, consequences of ongoing management practices on these species in France are unknown. For this study, we have drawn on existing long term monitoring programmes on breeding birds (French Breeding Bird Survey, FBBS) and trapping programmes in France to study the control efficiency on magpie, the most frequently species targeted by control activities (ROC, 2000). The magpie is viewed as a recurrent problem by conservationists and many hunters in France and Europe because it is predator of song birds and game birds (Mora, 2000; Birkhead, 1991). However, the impact of magpies on abundance and persistence of prey populations as well as control of its populations is controversial. Recent studies have suggested no correlation between magpie abundance and decline of its prey at local and national scales (Thomson et al., 1998; Chiron and Julliard, 2007; White et al., 2008; Newson et al., 2010). Information on the total number of individuals eliminated each year is scarce but in 2000, at least 402,000 magpies were killed in France, with wide variations between regions in the numbers killed (ONCFS, 2000; ROC, 2000). Trapping can have detrimental effects on magpie populations as it removes breeding birds (i.e. adult and immature birds) during the reproductive season. In France, there has been a steep decline in magpie numbers in the countryside (76% since 1990, Jiguet, 2010), while at the same time, magpies have colonised and established populations in urban environments (Chiron and Julliard, 2007). We suspect that the decline of magpie populations in France is due to trapping which is much more common in agricultural and natural environments than in cities (Chiron, 2007). But evidence of causal link between the decline of magpies in the countryside and trapping pressure is lacking. Magpies become established in cities thanks to their ability to exploit man-made resources and to the low predation rate on their nests (Jerzak, 2001). Whether or not cities have become a refuge for magpies because of the lack of hunting and trapping pressure is also unknown. We addressed these issues in a study of the spatial dynamics and demography of magpies in relation to human presence and management activities in France, using data on the occurrence and demography of magpies and control practices. Specifically, we assessed the impact of trapping activities on the age structure of local magpie populations. We estimated the age ratio between adults (i.e. more than one year) and immature (i.e. first year) to measure imbalances in breeding populations (Williams et al., 2002). The age ratio of breeding magpies is relevant because trapping activities target individuals that are reproducing when they are territorial, usually adults. By removing adults from breeding territories, trapping effort can attract immature in territories that

99

they were previously excluded. We predict an increase of the proportion of immature occupying breeding territories relatively to adults with trapping effort. We then investigated the dynamics of magpie occupancy, extinction and colonisation to test whether any effects on the age structure could be linked with changes in the persistence of populations at landscape and regional scale. We predict that trapping effort increased the probability of local extinction, and subsequently decreased the probability of magpie occupancy. We concluded with a discussion on the effects of control on magpie populations, management strategies, and the usefulness of bird surveys in dealing with the potential consequences of control activities. 2. Materials and methods 2.1. Estimating age structure We used data from a two-year national survey that we launched in 2003 to study the age structure of magpie populations. We asked volunteer trappers to send us the wings of magpies that had been killed between March and September during the breeding season. No magpies were killed for the purpose of this study and were trapped legally, following the appropriate guidelines for the human trapping and killing of the birds. To capture magpies, trappers use cages in which they place a live magpie to attract local breeding individuals (Díaz-Ruiz et al., 2010). Magpies typically start breeding as adults when they are two years old (Birkhead, 1991), but can sometimes reproduce when immature in their first year. We aged 98% of the wings received as adult, immature or young (i.e. year bird just fledged) depending on plumage characteristics: individual with small white patches on the tips of their feathers were aged as young (fresh feathers) or as immature (old feathers), adults (more than one year old) have large white patches on their feathers (Svensson, 1992). 2% Were undetermined because of unclear wing feathers pattern. As only breeding individuals are targeted by trappers and because of the purpose of the study, we removed young individuals from our sample. The age distribution was estimated in terms of the relative proportion of immature birds (PIs) captured in the breeding population (adults and immature). This measure is a composite index of both survival and recruitment of individuals into the breeding population and is suitable for assessing the impacts of control (Whitfield et al., 2007) as well as of hunting (Besnard et al. 2010; Miller and Otis, 2010). 2.2. Trapping effort As well as collecting wings, we asked trappers for additional information on (1) the number of traps they used in the field, (2) the number of days spent trapping per year, (3) the proportion of years where trapping activities was carried out in the last ten years, (4) the number of other trappers working the same location, and the day and location of capture. With this information, we developed an index to estimate the trapping effort as the product of (1), (2) and (3). We took (4) as an additional factor in the calculation, to estimate the Local Trapping Effort (LTE) as: LTE = (1)  (2)  (3) + (4)  [m (1)  m (2)  m (3)], where m (1), m (2) and m (3) are mean values per region. In addition to the LTE index, we compiled information from the regional offices in charge of species regulation (DDAF) on the total number of magpies killed by trappers per region in 2000 (ROC, 2000). Total number of magpies reported by regional offices correspond to magpies captured using the same trapping techniques as those used by trappers who participated to the ‘wing survey’ in 2003 and 2004. We obtained the total number of magpies killed for 53 of the 95 French regions (Fig. 1, information was unavailable

100

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

Fig. 1. Total number of magpies trapped in 53 regions in France for 2000 and distribution of the 2  2 km2 Breeding Bird Survey plots from 2001 to 2005 (n = 396, white circles).

2.4. Data on habitats and landscapes

in the other regions). We then calculated the Regional Intensity of Trapping (RIT) as the number of magpies killed divided by the number of breeding magpies per region. The number of breeding birds was derived from regional bird atlases (Chiron, 2007).

Magpie habitat: each point count for FBBS plots was classified by observers as belonging to semi-natural habitat, farmland or urban land. We selected plots with the same habitat type on at least 6 of the 10 sampling points of a plot. A total of 130 plots were thus classified as semi-natural habitats, 254 as farmland and 13 as urban land. Trappers were also asked to classify the habitat type surrounding each trapping location in 2003 and 2004 (n = 220) within a fixed 100 m radius as urban land, urban parks, farmland or seminatural habitat. Landscape scale: we used data from TERUTI, which is a statistical land cover survey, to quantify the proportion of urban land (PUL, e.g. parking, roads, railways, buildings, parks) in 2002 at each BBS plot (PULp) and within a 500 m radius around trapping locations (PULl) (ACDA, 2003). Since the TERUTI and the magpie survey locations were designed independently, we used the kriging interpolation technique to adapt the PUL measurements to the distribution of the magpie survey locations (Ashraf et al., 1997).

2.3. Data on magpie occupancy We used data from the French Breeding Bird Survey (FBBS), which has covered all French regions since 2001. The sampling method calls on skilled ornithologists who volunteer to count all bird species using a standardised procedure (Jiguet, 2010). On the same morning, each observer counts the birds (visible individuals and singers) at 10 fixed location points (at least 300 m apart) during 5-min intervals. Each point is visited twice a year (from April to mid-June). To be valid for temporal comparisons, the count must be repeated each year on approximately the same date (±7 days), at the same time of the day (±15 min within 1–4 h after sunrise) and by the same observer. For each given point, we used the occurrence of magpie from the two annual visits. The 10 fixed locations are evenly distributed within a randomly selected 2  2 km2 permanent plot (see Table 1). The random selection ensured that varied habitats were surveyed (including farmland, forest, suburbs and cities). Because the information on the RIT was available in 2000 only, we used plots of the FBBS between 2001 and 2005. A total of 397 FBBS plots were visited (3970 points), with each plot surveyed over 2–3 years on average between 2001 and 2005 (Fig. 1). Although the data on trapping activities (2000, 2003–2004) do not exactly overlap the magpie sampling period (2001–2005), we assume that the trapping effort was fairly constant throughout this period (Mora, 2000).

2.5. Modelling the effects of trapping on magpie spatial dynamics As magpie detection may vary in space and time, we used site occupancy models as developed by MacKenzie et al. (2003) to correct all estimates (occupancy, extinction and colonisation parameters) by magpie detection probability (see Appendix A). Following MacKenzie et al. (2003), the estimated actual occupancy (w2001) is the probability that a FBBS plot is occupied by magpies in 2001 and (pt) is the probability of detecting magpies given their presence on one sampling occasion at year t (i.e. a plot

Table 1 Magpie presence detected at the 2  2 km2 French Breeding Bird Survey plot level in main habitats between 2001 and 2005. Naive magpie occupancy was calculated as the ratio between detected presence and total number of plots; corrected occupancies are estimates from model 1 for 2001 (Table 2). Habitats

Species presence

2001

2002

2003

2004

2005

Total number of plots

Occupancy Naive (se)

Corrected (se)

Farmland

Non-detected Detected

7 45

35 106

37 141

32 125

38 163

149 580

0.80 (0.01)

0.88 (0.04)

Urban

Non-detected Detected

0 1

1 4

0 9

0 8

0 11

1 33

0.96 (0.01)

0.97 (0.13)

Semi-natural

Non-detected Detected

4 7

22 27

43 37

40 32

60 52

169 155

0.48 (0.01)

0.59 (0.09)

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

101

in 2001, 2002, etc.), (et) is the probability that a plot occupied in year t will become vacant in t + 1 (local extinction) and (ct) the probability that a vacant plot in year t will be occupied by the species in t + 1 (local colonisation). We tested three hypotheses for magpie occupancy as described in Appendix A. Occupancy 2001, extinction, colonisation and detection were modelled as a logistic function of the proportion of urban lands at plot level (PULp) and the regional intensity of trapping (RIT) and as a function of the discrete habitat variable. We examined more complex relationships between parameters and covariates using a continuous quadratic function that provided us with a flexible general model. We fitted a large number of models to the data and selected the best ones following a model selection procedure. We started modelling parameters with all predictors (full model) and removed each variable step by step in order to come out with a reduced model where all parameters are constant. The different models were then ranked using Akaike’s information criterion (AIC; Burnham and Anderson, 2002). The models with the lowest AIC, and AIC differences of less than 2, have a substantial level of empirical support (Burnham and Anderson, 2002). After the model selection, we plotted extinction, colonisation and occupancy parameters estimated based on the best model(s) against the variables ‘intensity of trapping’ and ‘proportion of urban areas’. Finally, we calculated equilibrium occupancy probabilities which are long-term occupancies if the effects of trapping are sustained and assuming constancy in parameter estimates. It was calculated as (c)/(e + c) with (e) the extinction probability and (c) the colonisation probability (Martin et al., 2009). Occupancy analyses were carried out with the PRESENCE program (Hines, 2006). 2.6. Modelling the effect of trapping on the proportion of immature birds For local analyses, we compared the proportion of immature birds captured by trappers according to habitat types and tested the relationship between the proportion of immature birds and the local trapping effort (LTE), taking the effects of the proportion of urban land (PULl) and the day of capture into account as control variables. We used mixed modelling and the lmer function in the package lme4 in R (Bates et al., 2011). Because magpies can be trapped several days consecutively at a same location, the day of trapping was used as a temporal replicate in the model. We coded the variable for the location as the random effect variable. We used a logit-link function assuming a binomial distribution of the proportion of immature birds. We used the R statistical software version 2.13.1 (R-Development Core Team, 2006). 3. Results 3.1. Relationship between the proportion of immature birds and the local trapping effort In 2003 and 2004, we received wings sent by trappers from 220 different localities in 34 regions. 1649 Magpies were aged as adults (n = 873, 53%) and immature (n = 776, 47%). Using the full dataset, we showed that the majority of wings were collected in urban and farm lands (n = 913 and 682, respectively) rather than in seminatural areas (n = 54). Relatively more adults were captured in urban areas than in farmlands and semi-natural habitats (Fig. 2a). Surprisingly, we found no statistical difference in the local trapping effort between farmlands, semi-natural and urban lands (LTEfarm = 5.37 ± 0.43 (se), LTEsemi-natural = 4.80 ± 0.72 (se), and LTEurban = 4.89 ± 0.26 (se), F = 1.39, df = 3, p-value = 0.26). The proportion of immature birds increased linearly with the local

Fig. 2. Variations in the proportion of immature birds captured (a) in three different habitats (for each habitat, n is the number of magpies sampled and % the proportion of immature), and (b) along a gradient of trapping intensity (each dot represents a trapping location, n = 143). Data were collected in 34 different French regions in 2003 and 2004. We tested differences in proportions between habitats, giving pvalues: p < 0.05, p < 0.01, p < 0.001.

trapping effort (z = 4.50, p-value < 0.001, Fig. 2b), but decreased linearly along an urbanisation gradient (z = 2.53, p-value = 0.011). Although not significant, the relationship between trapping intensity and the proportion of immature birds held true when using data on trapping at regional level (RIT) (rs = 0.55, df = 11, p-value = 0.08).

3.2. Relationship between spatial dynamics of magpies and regional trapping intensity Model selection indicated that occupancy (2001), extinction and colonisation at the FBBS plot level vary with the proportion of urban areas (PULp), the habitat type and the regional intensity of trapping (Models 1, 2 and 3, Table 2). Regional trapping intensity and the proportion of urban land (PULp) had additive but opposite effects on the probability of local extinction. The probability of local extinction was positively correlated with the regional intensity of trapping, while it decreased with the PULp (Fig. 3a). The model selection does not support an interaction effect between PULp and trapping intensity on the probability of extinction. However, we noted variations among habitats in the probability of local colonisation (Table 2). Derived from estimates obtained with model 1, the mean probability of colonising vacant territories was higher in urban areas (c = 0.99 ± 0.002 SE) than in farmlands (c = 0.38 ± 0.05 SE) and semi-natural landscapes (c = 0.06 ± 0.02 SE). Estimates obtained from the top 4 models suggest a positive relationship between PULp and the proportion of plots occupied

102

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

Table 2 Summary of model selection procedure results for occupancy (2001) and related parameters for the Magpie. We used data from the French Breeding Bird Survey covering the country from 2001 to 2005. The models are supported by biological hypotheses and ranked by D. No.

Modelsa

1 2 3 4 5 6 7 8 9 10

w(PUL) e(PUL + RIT) c(H) p(Hx PUL2) w(H + PUL) e(PUL + RIT) c(H) p(Hx PUL2) w(H + PUL + RIT2) e(PUL + RIT) c(H) p(Hx PUL2) w(PUL2) e(PUL + RIT) c(H) p(Hx PUL2) w(PUL) e(PUL2 + RIT) c(H) p(Hx PUL2) w(H + PUL + RIT) e(PUL + RIT) c(H) p(Hx PUL2) w(PUL) e(PUL + RIT2) c(H) p(Hx PUL2) w(H + PUL + RIT) e(RIT) c(H) p(Hx PUL2) w(H + PUL + RIT) e(PUL + RIT) c(H + PUL) p(Hx PUL2) w(H + PUL + RIT) e(PUL + RIT) c(H + PUL + RIT) p(Hx

11 12 13 14

w(H + PUL + RIT) e(PUL) c(H) p(Hx PUL2) w(H + PUL + RIT) e(PUL + RIT) c(PUL + RIT) p(Hx PUL2) w(H) e(PUL + RIT) c(H) p(Hx PUL2) w(H + PUL + RIT) e(PUL + RIT + year) c(H + PUL + RIT)

15

w(H + PUL + RIT) e(PUL + RIT) c(H + PUL + RIT)

w

np

Hypotheses on e and c variations

0.0 0.1 0.7 1.8 3.1 3.5 3.6 6.4 7.0 10.9

0.27 0.26 0.19 0.11 0.06 0.05 0.05 0.01 0.01 0.00

15 17 19 16 16 18 16 17 19 20

13.6 17.5 20.3 22.6

0.00 0.00 0.00 0.00

17 18 16 24

Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and of habitat and trapping on colonisation Effects of urbanisation on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and colonisation Effects of urbanisation and trapping on extinction and of habitat on colonisation Effects of urbanisation and trapping on extinction and colonisation

25.5

0.00

24

Effects of urbanisation and trapping on extinction and colonisation

28.8 29.3

0.00 0.00

26 24

Effects of urbanisation and trapping on extinction and colonisation Effects of urbanisation and trapping on extinction and colonisation

29.4

0.00

25

Effects of urbanisation and trapping on extinction and colonisation

62.0 212.4 282.3

0.00 0.00 0.00

31 9 3

Effects of urbanisation and trapping on extinction and colonization Static occupancy, no extinction and no colonization Extinction and colonization are random events

D

PUL2)

p(HxPUL2)

16 17

p(HxPUL2 + year) w(H + PUL + RIT) e(PULx RIT) c(Hx PULx RIT) p(HxPUL2)

w(H + PUL + RIT) e(PUL + RIT) c(H + PUL + RIT + year) p(HxPUL2)

18

w(H + PUL + RIT) e(PUL + RIT) c(HxPULx RIT)

19 20 21

p(HxPUL2) w(Hx PULx RIT) e(PULx RIT) c(HxPULx RIT) p(HxPUL2) w(PUL) p(HxPUL2) w(.) e(.) c(.) p(.)}

a Factors affecting occupancy (w), extinction (e), and colonisation (c) probabilities include: the proportion of urban land at plot level ‘PUL’, the habitat ‘H’ and the intensity of trapping ‘RIT’. Given are the relative difference in AICc values compared to the top-ranked model (D), AICc weights (w) and the number of parameters (np) in various models of magpie dynamics. We tested quadratic polynomial function (2) and ran season-dependent models (year). Interactions are indicated by the multiplication sign, while the plus sign denotes models with main effect only.

in 2001 by magpies (Table 2, Fig. 3a). There were also more plots occupied in urban areas than in farmlands and semi-natural habitats in 2001 (w = 0.97 ± 0.13 SE, w = 0.88 ± 0.04 SE, w = 0.43 ± 0.08 SE, respectively). Because occupancy (2001) varied among habitats (Table 2), we ran new model selections for semi-natural and farmland habitat types separately, using the same method. This enabled us to test whether the results held true when we analyzed single habitats (except the ‘urban lands’ class for which we received little data, n = 13 plots). In farmland habitats, the proportion of plots occupied in 2001 by magpies decreased steeply along a gradient of trapping intensity (RIT) (Fig. 3b, Table B.1 see Appendix B). Interestingly, the relationship between occupancy and trapping intensity was opposite to that observed between extinction and trapping intensity (Fig. 3b). Conversely, occupancy in 2001 was positively correlated with the proportion of urban lands. In semi-natural habitats, the model selection supports the hypothesis that the trapping effort affects occupancy within regions (Table B.2, see Appendix B), although there was high heterogeneity in occupancy (2001) between plots (Fig. 3c). As a consequence of extinction and colonisation dynamics, we estimated equilibrium occupancy probabilities along the gradient of regional trapping intensity. Equilibrium occupancy was lower than occupancy in 2001, ranging between 0.6 and 0.9 (Fig. 4). This indicates a continuing decline in occupancy after the study was carried out (Fig. 4).

4. Discussion Our results indicate that, among factors affecting magpies directly and indirectly, control activities are an important driver of population changes. Trapping has effects on the demography and distribution of populations beyond the scale of areas where

trapping takes place, at landscape (i.e. 2  2 km2 plot) and regional scales, and in various types of environments including cultivated lands and semi-natural areas.

4.1. Trapping effects on demography Trapping intensity increases the proportion of immature birds captured in breeding populations. Modifications in the age structure could arise as a consequence of recruitment and/or changes in the survival of breeding birds. In areas free of trapping, most territories are occupied by adults, and immature birds will not reproduce until they find a territory or replace adults that have died naturally. Suppressing breeding adult by trapping lowers competition for resources and territories, thus producing a larger proportion of immature birds recruited into the breeding population (Newton, 1993). Recruitment of immature individuals has already been shown to increase with hunting effort (Besnard et al., 2010). A younger overall age structure where trapping intensity is high may also indicate an increase in mortality, if we assume that mortality induced by trapping is additive. However, it is not known whether mortality due to trapping is compensated by a reduction in natural mortality among the individuals remaining in the population or by increased reproduction and survival of young individuals due to lower competition (i.e. compensatory mortality hypothesis, Williams et al., 2002). For the magpies, we expect that the compensation (vs. additive mortality) scenario is habitat dependent and arises in urban habitats more often, such as cities and suburban areas. In these environments, densities of magpies are higher than in less urbanised habitats like semi-natural areas and farmland (Jerzak, 2001) where it has already declined significantly (Jiguet, 2010). Although the trapping effort can be as important in urban and suburban areas as in the countryside, it has not resulted in the same demographic changes perhaps because of compensation

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

103

(a) All habitats

(b) Farmland habitats

(c) Semi-natural habitats

Fig. 3. Predicted relationships between local extinction, local occupancy (2001), intensity of trapping and the proportion of urban lands at the BBS plot level (a) for all habitats together, (b) for farmlands and (c) for semi-natural habitats in France. Each dot represents the estimated plot value. Each value is the estimate obtained with Model 1 in Table 2 for (a) and Models 1 in Tables B.1 and B.2 for (b) and (c) respectively (Appendix B).

mechanisms. In urban environments, magpies can have densitydependent habitat occupancy and demography (Fernández-Juricic, 2001; Jerzak, 2001), a prerequisite for compensation mechanisms to occur. Population models and field studies on hunting of a variety of species have demonstrated that due to compensation, harvest may not reduce abundance of spring-breeders (Boyce et al., 1999). In the countryside, we suspect that trapping is a major driver of demographic changes that has led to the population decline by the increase of breeder mortality. 4.2. Effects of trapping and urbanisation on spatial dynamics Our results suggest that the extinction rate of local magpie populations increases along a gradient of regional trapping intensity. Obviously, we did not expect the spatial dynamics of magpie populations at the plot scale to be perfectly stable, as populations naturally face local extinction and local colonisation events (e.g. due to stochastic processes and source-sink dynamics, Dias, 1996). But trapping activities strongly decrease the persistence of the magpie population over time by increasing local extinction in plots (i.e. 2  2 km2), with less possibility for re-colonising these sites

later on. Colonisation events did not compensate for site vacancy in areas with intense trapping, except in more urbanised areas where the colonisation rate was higher than in semi-natural areas or farmlands. This strengthens our hypothesis that compensatory mechanisms limit the impacts of trapping on magpie mortality and in cities. High magpie densities in urban lands allow rapid (i.e. within-year) replacements of individuals eliminated by trapping. Individuals recruited into the population may be local nonbreeding birds seeking vacant territories or may immigrate from neighbouring populations. As a consequence, species occupancy (2001) which is a parameter resulting from dynamic extinction and colonisation processes is unaffected by trapping in urban areas and decreases along a gradient of trapping intensity in farmland and semi-natural habitats. If we assume constancy in parameter estimates and trapping efforts, we predicted that magpie occupancies would have continued to decline in response to trapping after 2001. Although we do not know recent dynamics after 2005 to validate this prediction, it however suggests that trapping activity and its consequences must be viewed in the long term. Compared to trapping, urbanisation has a divergent positive effect on magpie persistence. The overall high occurrence of magpie

104

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

populations in cities is not due to the reduced trapping effort in this habitat but more likely result from the ability of magpies to colonise new man-made habitats and to exploit a wide range of food sources, including remnants of human food (Tatner, 1983). Additional food sources could influence breeding success, survival of chicks, juveniles, immature birds and adults (Hogstedt, 1981; Birkhead, 1991). For this study, we collected data over a short period, and it might be argued that differences between regions in magpie parameters represent short-term states for each population, perhaps related to changeable food resources, rather than illustrating long-term differences related to trapping pressure. We believe this interpretation is unlikely because it does not account for the differences in the distribution of populations along the gradient of control at different spatial scales. Additionally, the magnitude of the trapping intensities reported implies that trapping can be a major force in population dynamics (Williams et al., 2002). For estimating occupancy, we are aware of potential estimation bias due to the use of spatial subunits, instead of temporal replicates, especially when the species is absent from a subset of these spatial subunits (Kendall and White, 2009). We think this bias is unlikely because magpie home range is large enough and allows movements between locations within a plot during the sampling period such that territory range of a breeding pair can cover several points of a plot. Also, the magpie is a relatively common and conspicuous species. At plot level, probability of magpie detection was estimated ranging from 0.92 to 1 in a previous article (Chiron et al., 2008). According to Kendall et al. (2009), percentage of bias declined with increasing occupancy or detection probability. 4.3. Is ongoing control of magpies in France efficient? Trapping in France is probably a large component of total magpie mortality that has led to significant restriction in the magpie’s distribution range that is effective at local (2  2 km2 scale) and regional scales. These changes likely reflect current control activities effects as well as a history of control. Although efforts to control pest populations in France remain poorly documented, control has been broadly applied to many pest species since a long time (ROC, 2000). Past and current trapping efforts are likely responsible for the steep decline in magpie numbers observed in the French countryside since 1990 at least (Jiguet, 2010). Interestingly, consequences of control vary between habitats since resilience of populations to control activities was higher in urban areas than in the countryside. This means that a reduction of trapping effort would lead to some increase in the magpie population in countryside only. This population is below a density at which suppression of breeding individuals is extreme, and our evidence suggests that trapping is indeed at least partly additive to other forms of mortality. In urban areas, magpie populations regulate their own numbers which are not affected by control efforts but level of resources. Although the magpie is often targeted by trapping for reducing predation on game birds, the benefit of trapping on prey populations is uncertain. At local scale, magpie removal is ineffective to increase populations of common passerines (Chiron and Julliard, 2007; White et al., 2008). Also, there is little evidence that magpies reduce prey populations in the long term at large scale (Thomson et al., 1998). 4.4. National monitoring schemes as a new tool for studying largescale impacts of control With pest-related issues growing rather than diminishing, ecologists will need sufficient resources to maintain current research if they are to provide the understanding required to offer and

Fig. 4. Predicted equilibrium occupancy probabilities along a gradient of trapping intensity for farmland at the BBS plot level. Each dot represents the estimated plot value. Values were calculated as the ratio (c)/(e + c) with (e) the extinction probability and (c) the colonisation probability (Martin et al., 2009). Each value is given by the estimation obtained with Model 1 in Table B.1 (Appendix B).

evaluate sound management. Because nationwide monitoring cover a diversity of environmental contexts and geographical regions, such monitoring help understanding adequately the efficiency of population control, which is necessary to plan successful management program and to improve existing ones over large scale. Our study indicates that combining existing monitoring schemes is a relevant approach to assess large-scale impacts of control. This approach maximizes the use of available resources while maintaining the cost of evaluation relatively low. They also offer supplementary possibilities like testing the benefit of control activities on prey species directly using Breeding Bird Surveys (BBSs). Monitoring the age ratio of hunted and controlled populations are indirect methods that cannot be regarded as a substitute of capture recapture methods to study population dynamics. Such monitoring however enables a rapid assessment of survival of territorial species across populations and countries that could be maintained in the long term with the help of hunters and trappers to collect materials. 4.5. Conservation implications Variability in control efficiency between habitats indicates that the benefit expected from the control is not always proportional to the cost (i.e. the effort of control). This suggests that control activities must be viewed and planned in the light of target populations and habitat contexts. At countryside, trapping seems to be an efficient control method but we recommend it only if justified by management or conservation specific purposes and if another efficient non-lethal method could not be applied. Although trapping is the most direct approach, it has some limitations. When planned at region and country level, it is labour intensive and can be limited by the ability of some breeding individuals to learn to avoid traps (Alsager et al., 1972). At local scale, selective removal of magpies known to cause damages or occurring in areas with risk of damages could constitute a short-term means of reducing impacts. In case of high magpie density, removing territorial magpies may however increase risk of predation by non-territorial individuals in areas that they were previously excluded (Birkhead, 1991). In South of France, removing magpies may also endanger the Great spotted cuckoo (Clamator glandarius), a nest parasite of magpies (Soler and Møller, 1990). Non-lethal approaches of bird population management exist and could be more appropriate than trapping in the long term. Where magpies concentrate, like in urban parks, residential areas and other human settlements, we propose to reduce

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

sources of human-related food such as garbage dumps and feeders. Additional food sources influence life traits of individual birds (Hogstedt, 1981; Birkhead, 1991), nesting choice (Kang et al., 2012), and is likely to attract non-breeding birds (Chiron 2007). Non-lethal methods to prevent damages caused by corvids on seedlings, crops and fruits and on prey species in rural landscapes have also been tested successfully. For instance, taste aversion methods targeted at territorial ravens nesting in prey species habitats could reduce the attractiveness of prey species eggs and young (Cox et al., 2004). The effectiveness of such an approach over large areas and in the case of the magpie is however uncertain. Our evidence suggests that management activities should be better coordinated on a regional scale to optimise the effort and increase efficiency of measures. This is particular relevant in France and other European countries as management activities are locally applied but rarely regionally co-ordinated. Finally, the methodology used in this study could serve the understanding of trapping efficiency on other corvids such as the Rook, the Carrion crow and the Eurasian jay in France. Although less persecuted these three corvid species are also targeted by control measures (ROC, 2000). Acknowledgments We sincerely thank the hundreds of volunteers who took part in the national breeding bird survey and to whom this paper is dedicated (STOC EPS program), as well as trappers who actively participated to the national survey. This study was supported by the French ministry in charge of the environment, the Muséum National d’Histoire Naturelle and the Centre National pour la Recherche Scientifique. Thanks also to Audrey Muratet, two anonymous referees, David Miller for their helpful comments and to Ilona Bossanyi for editing the English version of the manuscript. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biocon.2012.08. 001. References Agreste Chiffres et Données Agriculture, 2003. Indicateurs paysagers élaborés à partir de l’enquête sur l’utilisation du territoire (TERUTI). In: Ministère de l’Agriculture de l’Alimentation de la Pêche et des Affaires rurales, vol. 151. Alsager, D.E., Stenrue, J.B., Boyles, R.L., 1972. Capturing Black-billed Magpies with circular live traps. J. Wildl. Manage. 36, 981–983. Ashraf, M., Loftis, J.C., Hubbard, K.G., 1997. Application of geostatistics to evaluate partial weather station networks. Agric. Forest Meteorol. 84, 255–271. Baker, P.J., Harris, S., 2006. Does culling reduce fox (Vulpes vulpes) density in commercial forests in Wales, UK. Eur. J. Wildl. Res. 52, 99–108. Bates, D., Maechler, M., Bolker, B., 2011. Lme4: Linear Mixed-effects Models Using S4 Classes. R Package Version 0.999375-41. . Beja, P., Gordinho, L., Reino, L.F., Loureiro, Santos-Reis, M., Borralho, R., 2009. Predator abundance in relation to small game management in southern Portugal: conservation implications. Eur. J. Wildl. Res. 55, 227–238. Besnard, A., Novoa, C., Gimenez, O., 2010. Hunting impact on the population dynamics of Pyrenean grey partridge Perdix perdix hispaniensis. Wildl. Biol. 16, 135–143. Birkhead, T., 1991. The Magpies. T & AD Poyser, London. Boyce, M.S., Sinclair, A.R.E., White, G.C., 1999. Seasonal compensation of predation and harvesting. Oikos 87, 419–426. Burnham, K.P., Anderson, D.R., 2002. Model Selection and Inference. A Practical Information Theoretic Approach. Springer-Verlag, New York, USA. Chiron, F., 2007. Dynamiques spatiale et démographique de la pie bavarde Pica pica en France: implications pour la gestion. PhD thesis, Muséum national d’Histoire naturelle. Chiron, F., Julliard, R., 2007. Responses of songbirds to Magpie reduction in an urban habitat. J. Wildl. Manage. 71, 2624–2631. Chiron, F., Lee, A., Julliard, R., 2008. Effects of landscape urbanization on magpie occupancy dynamics in France. Landscape Ecol. 23, 527–538.

105

Conover, J., 2002. Resolving Human-Wildlife Conflicts: The Science of Wildlife Damage Management. Michael. CRC Press, Boca Raton, FL, USA, 2002. xiv. Côté, I.M., Sutherland, W.J., 1997. The effectiveness of removing predators to protect bird populations. Conserv. Biol. 11, 395–405. Cox, R., Baker, S.E., Macdonald, D.W., Berdoy, M., 2004. Protecting egg prey from carrion crows: the potential of aversive conditioning. Appl. Anim. Behav. Sci. 87, 325–342. Devictor, V., Julliard, R., Couvet, D., Jiguet, F., 2008. Birds are tracking climate warming, but not fast enough. Proc. Royal Soc. B 275, 2743–2748. Dias, P.C., 1996. Sources and sinks in population biology. Trends Ecol. Evol. 11, 326– 330. Díaz-Ruiz, F., García, J.T., Pérez-Rodríguez, L., Ferreras, P., 2010. Experimental evaluation of live cage-traps for black-billed magpies Pica pica management in Spain. Eur. J. Wildl. Res. 56, 239–248. Fernández-Juricic, E., 2001. Density dependent habitat selection of corridors in a fragmented landscape. Ibis 143, 278–287. Gregory, R.D., Baillie, S.R., 1998. Large-scale habitat use of some declining British birds. J. Appl. Ecol. 35, 785–799. Gregory, R.D., van Strien, D.A., Vorisek, J., Gmelig Meyling, D., Adriaan, W., Noble, G., Foppen, R., Gibbons, D.W., 2005. Developing indicators for European birds. Philos. Trans. Royal Soc. B 360, 269–288. Harding, E.K., Doak, D.F., Albertson, J.D., 2002. Evaluating the effectiveness of predator control: the non-native red fox as a case study. Conserv. Biol. 15, 1114–1122. Heydon, M.J., Reynolds, J.C., 2000. Fox (Vulpes vulpes) management in three contrasting regions of Britain, in relation to agricultural and sporting interests. J. Zool. 251, 237–252. Hines, J.E., 2006. PRESENCE2. Software to Estimate Patch Occupancy and Related Parameters USGS. PWRC. . Hogstedt, G., 1981. Effect of additional food on reproductive success in the magpie (Pica pica). J. Anim. Ecol. 50, 219–229. Jerzak, L., 2001. Synurbanization of the magpie in the Palearctic. In: Marzluff, J.M., Bowman, R., Donnelly, R. (Eds.), Avian Ecology and Conservation in An Urbanizing World. Kluwer Academic, Norwell, Massachusetts. Jiguet, F., 2010. Les résultats nationaux du programme STOC de 1989 à, 2009, . Julliard, R., Jiguet, F., Couvet, D., 2004. Evidence for the impact of global warming on the long-term population dynamics of common birds. Proc. Royal Soc. Lond. B 271, S490–S492. Kang, W., Lee, D., Park, C.R., 2012. Nest distribution of magpies Pica pica sericea as related to habitat connectivity in an urban environment. Landscape Urban Plann 104, 212–219. Kendall, W.L., White, G.C.A., 2009. Cautionary note on substituting spatial subunits for repeated temporal sampling in studies of site occupancy. J. Appl. Ecol. 46, 1182–1188. McDonald, R.A., Harris, S., 2002. Population ecology of stoats Mustela erminea and weasels Mustela nivalis on game estates in Great Britain. J. Appl. Ecol. 39, 793– 805. MacKenzie, D.I., Nichols, J.D., Hines, J.E., Knutson, M.G., Franklin, A.B., 2003. Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84, 2200–2207. Martin, J., Nichols, J.D., Mc Intyre, C.L., Ferraz, G., Hines, J.E., 2009. Perturbation analysis for site occupancy dynamics. Ecology 90, 10–16. Miller, D.A., Otis, D.L., 2010. Calibrating recruitment estimates for mourning doves from harvest age ratios. J. Wildl. Manage. 74, 1070–1078. Mora, I., 2000. Nuisance et régulation des espèces animales susceptibles d’être classées nuisibles en Ile-de-France: descriptif de la situation. In: Paris, Conseil Régional Cynégétique de l’Ile de France. Newson, S.E., Evans, K.L., Noble, D.G., Greenwood, J.J.D., Gaston, K.J., 2008. Use of distance sampling to improve estimates of national population sizes for common and widespread breeding birds in the UK. J. Appl. Ecol. 45, 1330–1338. Newson, S.E., Rexstad, E.A., Baillie, S.R., Buckland, S.T., Aebischer, N.J., 2010. Population change of avian predators and grey squirrels in England: is there evidence for an impact on avian prey populations? J. Appl. Ecol. 47, 244–252. Newton, I., 1993. Predation and limitation of bird numbers. Curr. Ornithol. 11, 143– 198. Novaro, A.J., Funes, M.C., Walker, M.C., 2005. An empirical test of source-sink dynamics induced by hunting. J. Appl. Ecol. 42, 910–920. Office National de la Chasse et de la Faune Sauvage, 2000. Enquête nationale sur les tableaux de chasse à tir. Saison 1998–1999. Faune sauvage, cahiers, techniques, 251. Ormerod, S.J., 2002. Applied issues with predators and predation: editor’s introduction. J. Appl. Ecol. 39, 181–188. Péron, G., Ferrand, Y., Choquet, R., Pradel, R., Gossmann, F., Bastat, C., Guénézan, M., Bauthian, I., Julliard, R., Gimenez, O., 2012. Spatial heterogeneity in mortality and its impact on the population dynamics of Eurasian woodcocks. Popul. Ecol. 54, 305–312. R-Development Core Team, 2006. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. (accessed September, 2011). Rassemblement des Opposants à la Chasse, 2000. Tout savoir sur le piégeage. Réglementation, résultats d’enquête, analyse de revues. Ligue ROC, Paris. Rushton, S., Shirley, M., MacDonald, D., Reynolds, J., 2006. Effects of culling fox populations at the landscape scale: a spatially explicit population modelling approach. J. Wildl. Manage. 70, 1102–1110.

106

F. Chiron, R. Julliard / Biological Conservation 158 (2013) 98–106

Servanty, S., Gaillard, J.M., Ronchi, F., Focardi, S., Baubet, E., Gimenez, O., 2011. Influence of harvesting pressure on demographic tactics: implications for wildlife management. J. Appl. Ecol. 48, 835–843. Soler, M., Møller, A.P., 1990. Duration of sympatry and coevolution between the great spotted cuckoo and its magpie host. Nature 343, 748–750 (London). Svensson, L., 1992. Identification Guide to European Passerines. Marstatryck, Stockholm. Tatner, P., 1983. The diet of urban magpies. Ibis 125, 90–107. Thomson, D.L., Green, R.E., Gregory, R.D., Baillie, S.R., 1998. The widespread declines of songbirds in rural Britain do not correlate with the spread of their avian predators. Proc. Royal Soc. Lond. B 265, 2057–2062. Treves, A., 2009. Hunting for large carnivore conservation. J. Appl. Ecol. 46, 1350– 1356. Tryjanowski, P., Sparks, T.H., Kamieniarz, R., Panek, M., 2009. The relationship between hunting methods and sex, age and body weight in a non-trophy animal, the red fox. Wildl. Res. 36, 106–109.

Villafuerte, R., Viñuela, J., Blanco, J.C., 1998. Extensive predator persecution caused by population crash in a game species: the case of red kites and rabbits in Spain. Biol. Conserv. 84, 181–188. Virgós, E.S., Travaini, A., 2005. Relationship between small game hunting and carnivore diversity in central Spain. Biodivers. Conserv. 14, 3475–3486. White, P.J.C., Stoate, C., Szczur, J., Norris, K., 2008. Investigating the effects of predator removal and habitat management on nest success and breeding population size of a farmland passerine: a case study. Ibis 150, 178–190. Whitfield, D.P., Fielding, A.H., Mcleod, D.R.A., Morton, K., Stirling-Aird, P., Eaton, M.A., 2007. Factors constraining the distribution of Golden Eagles Aquila chrysaetos in Scotland: Capsule between 1992 and 2003 persecution appeared to be the main influential factor. Bird Study 54, 199–211. Williams, B.K., Nichols, J.D., Conroy, M.J., 2002. Analysis and Management of Animal Populations. Academic Press, San Diego, California, USA. Zipkin, E.F., Kraft, C.E., Cooch, E.G., Sullivan, P.J., 2009. When can efforts to control nuisance and invasive species backfire? Ecol. Appl. 19, 1585–1595.