SEARCH

SEARCH BY CITATION

Keywords:

  • ecological traps;
  • nest placement;
  • settlement cues;
  • predation;
  • Motacilla flava

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Human-induced habitat changes often generate novel or radically altered habitat characteristics, which can impair the ability of organisms to differentiate between suitable and unsuitable sites. This phenomenon, often termed an ecological trap, has been identified as a potential driver of biodiversity loss worldwide. However, few unequivocal examples have been documented, even in agricultural environments where contemporary habitat changes have been rapid and significant. Several problems complicate the detection of ecological traps in the field, including difficulties in measuring key parameters such as relative habitat preference. Here, we assess habitat selection preferences and breeding success of the yellow wagtail Motacilla flava, a UK red-listed declining passerine, in arable farmland. We combine habitat-specific density indices with measures of home range exclusivity to make inferences on relative habitat preference that are robust to the confounding effect of competitive exclusion. Using multiple measures of breeding success, we identify maladaptive habitat selection patterns at the scale of both territory and nest site choice. Yellow wagtails showed a preference for establishing territories within field bean crops, but subsequently suffered high nest predation rates. Similarly, pairs showed a preference for nesting close to tramlines within cereal fields, but nests further from tramlines achieved higher success due to lower predation rates. We found no evidence of competitive exclusion among neighbouring pairs, suggesting that density-based indices provided an accurate reflection of relative habitat preferences. Our findings highlight the potential role of maladaptive habitat selection in suppressing breeding success among farmland species.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In human-altered environments, changes in the appearance or functioning of habitats can lead to a disconnection between the habitat selection decisions made by organisms and underlying patterns of habitat quality (Schlaepfer, Runge & Sherman, 2002; Battin, 2004; Robertson & Hutto, 2006). In some circumstances, this can result in preferential settlement in poor quality ‘sink’ habitats, even when higher quality alternatives are available, a scenario often termed an ecological trap (Gates & Gysel, 1978; Schlaepfer et al., 2002; Battin, 2004). Studies reporting ecological trap phenomena have proliferated in recent years, with examples involving a wide range of taxa in a variety of habitats including grass prairies (Manzer & Hannon, 2005; Shochat et al., 2005), wetlands (Hernandez, Reece & McIntyre, 2006), rivers (Pelicice & Agostinho, 2008) and woodlands (Mänd et al., 2005). Surprisingly, relatively few examples have been described from intensive agricultural ecosystems (e.g. Rodenhouse & Best, 1983; Purcell & Verner, 1998), despite the rapid and significant changes that have occurred within these habitats in recent years (Newton, 2004; Gilroy & Sutherland, 2007). The detection of ‘true’ ecological traps (those that drive population decline) in field settings is often complicated by difficulties in measuring habitat-specific population growth rates (Battin, 2004; Pärt, Arlt & Villard, 2007). Rather, studies tend to focus on assessing whether settlers actively prefer habitats where per capita fitness is lower than that attainable in alternatives (i.e. settlement is maladaptive), with such patterns being taken to imply the presence of a trap (Battin, 2004; Robertson & Hutto, 2006). Many such studies have been criticized for the methods used to infer habitat preference, particularly the use of population density as an indicator of habitat selection decisions (Robertson & Hutto, 2006; Arlt & Pärt, 2007). For example, density-based indices can be misleading in circumstances where strong competitors exclude weaker ones from preferred habitat patches, forcing them to co-exist at higher densities in less preferred areas (Van Horne, 1983; Railsback Stauffer & Harvey, 2003; Johnson, 2007). Given this confounding effect, various authors have promoted the use of individual-based preference measures such as settlement order or sequential patch occupancy between years (Robertson & Hutto, 2006; Arlt & Pärt, 2007). Unfortunately, individual-based habitat preference measures are difficult to implement in many situations. Settlement order, for example, may be difficult to assess unless the study species is migratory (e.g. Remes, 2003), while measures of sequential patch occupancy may be logistically difficult or confounded by the effect of site fidelity (Johnson, 2007). The lack of widely applicable methods to assess relative habitat preference currently restricts the capacity for researchers to assess the consequences of habitat selection in many situations.

Here, we study habitat selection and breeding success in a ground-nesting passerine, the yellow wagtail Motacilla flava, in arable farmland. Like many farmland species, this species is undergoing a dramatic population decline in the UK, with numbers falling by 65% between 1970 and 2005 (Eaton et al., 2007). The causes of decline remain unclear, although changes in the management of grassland breeding habitats might be important (Wilson & Vickery, 2004), as well as the effects of soil degradation on prey availability (Gilroy et al., 2008). We hypothesized that recent changes in habitat appearance (associated with arable intensification) could have led to mismatches between habitat selection and fitness in this species, contributing to population declines. We use a novel method combining density-based habitat selection indices with measures of home range exclusivity to determine relative habitat preferences for yellow wagtails within arable farmland. Given the extent of variation across taxa in the magnitude of territorial home-range defence (Brown & Orians, 1970; Maher & Lott, 2000), the influence of competitive exclusion on population spacing varies widely between study systems (Cody, 1981; Kaufmann, 1983). By measuring the magnitude of competitive exclusion within the study population, we can assess the extent to which habitat-specific territory densities genuinely reflect relative habitat preferences. We relate observed habitat preferences to several measures of breeding success to assess whether habitat selection by yellow wagtails is adaptive within the study environment.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study areas

We surveyed six areas of exclusively arable farmland in Lincolnshire and Cambridgeshire, UK covering 33 km2 across 14 different farms. Three areas were surveyed in both 2003 and 2005, with a further three areas surveyed in 2005 only (see Table 1). All farms were under conventional intensive management, with cropped habitats occupying over 80% of the total area on each site. Broad landscape features were similar across all sites, with limited woodland (<10% of total area on all sites) and some cropped land set-aside from production under EU regulations (<15% of total area, Table 1). All areas were topographically similar, lying on an alluvial floodplain ranging from 0.5 to 5 m above sea level.

Table 1.  Characteristics of six study sites in the Fenland region of Lincolnsire and Cambridgeshire, UK
SiteSurvey yearsArea surveyed (km2)Field size (ha) (mean ± sd)Area (ha) of:
WheatField beansPeasPotatoesSugar beetOilseed rape
Archer's Drove20055.217.6 ± 6.522998000105
Borough Fen2003, 20054.412.2 ± 5.51293640479036
Bourne South Fen20056.112.6 ± 4.3186012598640
Deeping Fen2003, 20054.522.1 ± 7.921580007530
Dunsby Fen20059.215.2 ± 6.137216512000170
Langtoft Fen2003, 20053.913.3 ± 4.21425070224060

Territory mapping and nest monitoring

We surveyed each site twice per month during the breeding season (April–August), using territory mapping methodology (Marchant, 1983). Census routes followed field boundaries and tramlines such that observers passed within 50 m of each point within each field on each visit. Male advertisement songs and song-flight displays were used to identify territories, together with mobbing behaviour performed by both sexes around active nests (Smith, 1950; Roselaar, 1988). Once a territory was found, we carried out detailed observations of adult movements to identify areas of potential nesting activity, followed by exhaustive nest searches. In territories where initial searches failed to reveal a nest, we made repeat searches throughout the season to confirm that subsequent breeding did not take place. After discovery, we visited nests at 3-day intervals, recording clutch or brood size on each visit. Nests that were emptied before fledging (i.e. <12 days after hatching) were assumed to have been predated. The date of failure (or fledging) was estimated as the mid-point between the last visit when the nest was active and the date on which the outcome was known (Mayfield, 1975).

We carried out focal watches at each active nest during chick rearing (unless nest failure occurred before a focal watch could be conducted) in order to assess foraging range overlap between nesting pairs. Observers were concealed in habitats unsuitable for foraging (or within a vehicle if surrounding habitats were too open to allow concealment) at a minimum distance of 100 m from the focal nest, with each nest being monitored continuously for 1 h. We were not able to identify birds individually within our study area (very few were colour marked), so we assigned foraging movements to nests by recording foraging events only when they occurred immediately after observed nest visits. Adults were followed for the first flight after leaving the nest using binoculars, recording the initial landing location on 1:2500 maps. Aggressive interactions were also recorded. Foraging watches were carried out between 08:00 and 16:00 h in clear weather with winds <15 mph. In order to ensure that a similar number of foraging events was mapped for each nest, we repeated focal watches if fewer than five events were recorded in the first hour (13 of 39 observed nests in total). We mapped foraging destinations around each nest as straight radial lines from the nest site to the initial landing point using MapInfo (V. 6.5) gis software. We calculated the rate of intersection between the flight paths of simultaneously active pairs as an index of home range overlap.

Habitat variables

Sites were mapped using MapInfo on a field-by-field basis, with field size and boundary length calculated from 1:2500 maps. On each visit to each field, we measured crop height to the nearest centimetre at 10 random points using a 2 m rule, and estimated ground-level visibility at each point by counting visible grid intersections on a 20 × 30 cm white card marked with a 5 × 5 cm grid, placed perpendicular to the ground at a distance of 1 m from the observer. At nest sites, we measured crop height and density (proportion of bare ground in a 1 m2 quadrat) on each nest visit. We then back-calculated the approximate crop height the first-egg date of the nest assuming a linear growth rate of the crop. We measured the linear distance to nearest tramline using a 2 m rule, and distance to the nearest field by pacing.

Statistical analysis

We assessed relative preferences for field level territory choice and within-field nest placement using the Manly preference index a (Manly et al., 1972), a readily interpretable measure that is robust to differences in the available area of each habitat at different sites (Sutherland & Green, 2004), calculated thus

  • image

where r1 is the proportion of territories or nests in a given habitat category and n1 is the proportion of area covered by that habitat category, with n total categories. Values range between 0 and 1, with 1 indicating strong relative preference and 0 indicating strong relative avoidance. We calculated field-level preferences using the relative frequency of territories, rather than nests, as territory surveys were less likely to overlook individual breeding attempts. For within-field preferences, we used the relative frequency of nests in discrete habitat categories, for example distance bands from a field edge, such that each category had a similar sample size. Inferences on within-field preference therefore assumed equal likelihood of nest discovery across categories. Given that exhaustive nest searches were carried out in each territory, we believe this assumption was not violated. Preference index calculations were made separately for each study site as the calculation assumes a freedom of settlement choice between all habitat categories.

To determine the fitness consequences of habitat selection decisions, we examined variation in three measures of breeding success: clutch size, brood size at fledging (of non-predated nests) and nest predation likelihood, using generalized linear models. As yellow wagtails disperse and establish new territories (and potentially re-form pairs) between consecutive breeding attempts (Gilroy et al., 2010a), we considered all nests as independent with respect to habitat variation, despite the possibility that some were repeat attempts made by the same adults. Clutch size and non-predated brood size were modelled with Gaussian errors and identity link function, while predation likelihood was modelled under a logistic exposure approach with binomial error distribution and logit link function (Aebischer, 1999). For predation likelihood models, the response variable was nest predation (1) or survival (0), with the number of nest exposure days between detection and outcome as the binomial denominator (Aebischer, 1999). Initially, we tested for crop-level differences in each fitness component by assessing the change in residual deviance after the removal of a categorical crop term from a model incorporating fixed effects for site, year and date of nest initiation, via a likelihood ratio test (Crawley, 1993). Crop-level variation in predation likelihood was modelled for egg and chick stages separately.

In order to examine within-field predictors of each fitness component, we used an information–theoretic approach to evaluate the performance of four predictor variables: (1) linear distance from nest to tramline; (2) linear distance from nest to field edge; (3) height of the crop over the nest; (4) crop density around the nest. We modelled data from each crop type separately as within-field patterns were deemed likely to be crop specific; crops with very small sample sizes (<20 nests) were excluded from these analyses. For each response variable, we constructed models from all possible linear combinations of the four variables (with fixed effects for year, site and nest initiation date), performing model selection with AICc (Burnham & Anderson, 2002). We ranked models based on AIC weights (AICw) and used model averaging to estimate parameters and unconditional standard errors from the full candidate set. The likelihood that each predictor should be included in the best model was estimated using AIC selection probabilities, calculated by summing AICw of all models containing each variable (Burnham & Anderson, 2002). For within-field predation likelihood models, we pooled data from egg and chick stages in order to maximize crop-specific sample sizes. In these cases, parameter estimates represent indices of predation likelihood in relation to habitat variables, rather than absolute measures of predation rate. All statistical analyses were conducted using the r statistical package version 2.0.1 (R Project 2006).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We monitored 137 active yellow wagtail nests in total. Most nests successfully fledged at least one offspring (63.5% of total, n=87), and among nests that failed (n=50), predation was the most frequent cause of failure (n=42, 30.7% of all nests). We found almost complete overlap in the home ranges of simultaneously active nests (n=39 observed pairs, Fig. 1). Of all recorded foraging paths (n=638, mean per nest=16.36 ± 0.55 se), 41% intersected at least one path of another simultaneously active pair (mean intersections per path=1.24 ± 0.41 se). Only five aggressive interactions were noted during all focal watches, and individuals were frequently observed foraging in close proximity to those from neighbouring pairs without interference. These observations suggest that competitive home-range defence was weak within our study area, and hence that territory density was likely to be a robust indicator of relative habitat preference.

image

Figure 1.  Observed foraging locations and nest sites of yellow wagtail Motacilla flava pairs within a section of one study site (Bourne South Fen, Lincolnshire, UK). Lines represent field edges; letters a–k correspond to individual nests and show locations visited by each pair during focal observations; encircled letters indicate nest sites; encircled ‘n’ indicates nests for which home range data could not be collected (due to, e.g. early failure).

Download figure to PowerPoint

Crop-level patterns of habitat preference and nest success

Nests were found in four crops: wheat (autumn sown), field beans, field peas and potatoes (all spring sown). The distribution of territories indicated that relative preference was highest for field bean crops (a=0.55 ± 0.20) and potatoes (a=0.54 ± 0.02), with lower relative preference for wheat crops (a=0.29 ± 0.13) and peas (a=0.20 ± 0.05). Likelihood ratio tests showed no significant effects of crop type on clutch size (χ2=3.7, d.f.=3, P=0.31) or brood size at fledging (χ2=2.2, d.f.=3, P=0.49). Nest predation likelihood, however, varied significantly in relation to crop type during both the chick stage (χ2=16.3, d.f.=3, P<0.001), and egg stage (χ2=4.2, d.f.=3, P=0.021). The pattern was similar across both nest stages (Table 2), with nests in field bean crops having the highest overall predation likelihood (0.91 ± 0.08, n=15), with the remaining crops having predation likelihoods varying from 0.58 to 0.70. Figure 2 shows the influence of crop type on overall predation likelihood, together with the preference ranking of each crop (Manly et al., 1972). Non-predated brood sizes at fledging were similar between field beans and other crops (field beans mean=4.55 ± 0.2; other crops mean=4.59 ± 0.2), suggesting that the increased predation risk was not buffered by high nest productivity. Ground level visibility in field bean crops increased with crop growth, caused by the abscission of lower leaves following increased shading from the upper canopy (Fig. 3). This did not occur in any of the other crops selected for nesting by yellow wagtails.

Table 2.  Estimates of predation likelihood (and standard errors) for yellow wagtail Motacilla flava pairs nesting in the four main crop types used in the study region (Lincolnshire and Cambridgeshire, UK)
CropPredation probability
Egg stagenChick stagenWhole periodn
  1. Sample sizes (n) indicate number of observation days for nests in each crop, at each stage, with the number of nests contributing in parentheses.

Autumn wheat0.668 ± 0.02191(27)0.167 ± 0.01330 (51)0.587 ± 0.02521 (63)
Potatoes0.724 ± 0.02137(24)0.325 ± 0.01341 (41)0.603 ± 0.02478 (48)
Field beans0.963 ± 0.1017 (4)0.569 ± 0.0374 (13)0.919 ± 0.0891 (15)
Peas0.622 ± 0.0618 (4)0.427 ± 0.0344 (8)0.701 ± 0.0862 (9)
image

Figure 2.  Preference indices (a) and nest predation likelihoods (b) for the four crop types in which sufficient samples of yellow wagtail Motacilla flava nests were found, in six areas of arable farmland in eastern England in 2005. Manly's preference index (± 95% confidence interval) is a relative measure of use (territory abundance) in relation to availability of each crop. Values close to 1 indicate strong preference, while values close to 0 indicate strong avoidance. Sample sizes (n) indicate number of nests monitored in each crop.

Download figure to PowerPoint

image

Figure 3.  The relationship between crop height and horizontal visibility at ground level in pea, potato, autumn-sown wheat and field bean crops. Visibility index is the number of grid intersections visible on a 20 × 30 cm piece of white card (5 × 5 cm grid) placed perpendicular to the ground, at a distance of 1 m from the observer.

Download figure to PowerPoint

Within-field patterns of habitat preference and nest success

Sample sizes were sufficient (>20 nests) to explore within-field patterns of nest placement and breeding success in autumn-sown wheat and potato crops only. In wheat crops, yellow wagtails showed a strong preference for placing nests close to tramlines (Fig. 4a), with 62% of nests being placed within 25 cm of a tramline. The crop interior (areas exceeding 3 m from the tramline) was strongly avoided for nest placement (a=0.01 ± 0.004). In other crops, nest placement showed no clear pattern with respect to tramlines. In both wheat and potato crops, bird strongly avoided nesting <60 m from a field edge (a=0.05 ± 0.06 in potatoes and 0.03 ± 0.04 in wheat), showing a relative preference for areas exceeding 100 m from field edges in both crops (Fig. 5a and c).

image

Figure 4.  Preference indices (a) and nest predation likelihoods (b) for yellow wagtails Motacilla flava nesting in autumn-sown wheat crops at different distances from the nearest tramline. Preference index values close to 1 indicate strong preference, while values close to 0 indicate strong avoidance. Sample sizes (n) indicate number of nests monitored in each distance band.

Download figure to PowerPoint

image

Figure 5.  The effect of distance from the nearest field edge on nest-site preference and likelihood of nest predation in potato (a and b) and autumn-sown wheat (c and d) crops. Preference index values close to 1 indicate strong preference, while values close to 0.5 indicate use in proportion to availability and values close to 0 indicate avoidance. Sample sizes (n) indicate number of nests monitored in each distance band. Right-hand data points indicate all areas exceeding 100 m (a and b) or 150 m (c and d).

Download figure to PowerPoint

None of the four within-field habitat variables had significant power in explaining variation in clutch or brood size in either crop type (AIC selection probabilities<0.75 in each case, Table 3). In contrast, we found significant variation in nest predation likelihood with respect to within-field habitat features in both crops. In wheat, the best model (AICw=0.347) included distance to nearest tramline and distance to nearest field edge, both of which had high AIC selection probabilities (0.95 and 0.96, respectively). Nest predation likelihood increased significantly with proximity to tramlines; for nests placed within 10 cm of a tramline, the mean nest predation likelihood was 0.75 (n=27; Fig. 4b), while we recorded no predation events among nests placed more than 3 m from a tramline (n=15). In potato crops, the best model (AICw=0.284) included only distance to nearest field edge (AIC selection probability 0.95, Table 3). This relationship was consistent across both crops (Fig. 5), with no predation events occurring among nests placed at least 150 m from field edges (n=9 for potatoes; n=16 for wheat). Vegetation height and density had low selection probabilities for all response variables, suggesting limited influence on breeding success (Table 3).

Table 3.  Selection probabilities, model-averaged coefficients and unconditional standard errors (se) for all within-field variables modelled to explain variation in three components of nest success: clutch size, non-predated brood size and nest predation likelihood
Fitness componentPredictorWheatPotatoes
Selection probabilityEstimateseSelection probabilityEstimatese
  1. Selection probabilities are summed AICw values for all models that included each given variable, representing a probabilistic indicator of the plausibility of the effect.

  2. AICw, AIC weight.

Clutch size
 Distance from tramline0.374−0.00020.0010.3890.00010.001
Distance from field edge0.413−0.00010.0010.542−0.00030.001
Crop density0.3930.03320.0540.6880.05570.082
Crop height0.3680.00040.0010.714−0.00420.002
Brood size
 Distance from tramline0.3180.00020.0080.3700.00010.001
Distance from field edge0.3570.00010.0010.542−0.00030.003
Crop density0.4120.8901.0350.3950.12200.184
Crop height0.7120.0270.0290.391−0.00260.003
Nest predation likelihood
 Distance from tramline0.9450.00170.0010.3710.00000.001
Distance from field edge0.9610.00120.0010.9540.00190.001
Crop density0.4220.00180.0020.386−0.00070.001
Crop height0.410−0.00130.0010.3130.00020.001

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Yellow wagtails showed two patterns of maladaptive habitat selection in our study area, namely preference for field bean crops and proximity to tramlines in wheat, both of which resulted in increased nest predation risk relative to less-preferred habitats. We found no evidence that these fitness costs were buffered by increases in other fitness components (either clutch or brood size). One other habitat feature, proximity to field boundaries, was negatively correlated with predation risk and positively with habitat choice, suggesting adaptive habitat selection in relation to this feature (Morris & Gilroy, 2008).

Breeding pairs showed a high degree of simultaneous home-range overlap with minimal aggressive interaction, suggesting that the spacing of territories was unlikely to be influenced by competitive exclusion within our study area (Cody, 1981; Maher & Lott, 2000). Consequently, we were able to use simple density-based indices to assess relative habitat preference without recourse to resource-intensive individual-based preference measures (Robertson & Hutto, 2006; Arlt & Pärt, 2007). While the criticisms levied at density-based habitat preference indices (e.g. Robertson & Hutto, 2006) are clearly valid in many circumstances, our results demonstrate that they should not be regarded as a blanket restriction. Indeed, we suspect that our approach, based on combining density-based indices with direct measurement of confounding effects, may be applicable in a wide range of situations where individual-based preference measures are impractical. Given the ubiquitous importance of habitat preference assessment in conservation biology, we suggest that our approach will help widen the range of tools available to measure habitat selection in challenging field conditions.

In yellow wagtails, preferential settlement in field bean crops appeared to relate to the suitability of the crop sward structure in the early stages of growth. A concomitant study of foraging habitat selection in the study area (Gilroy et al., 2010b) found that bean crops were relatively low-ranked in terms of foraging preference, suggesting that territory settlement in beans was unlikely to be due to high foraging success. Rapid changes in vegetative structure during crop growth (principally the loss of lower leaves) caused dramatic increases in ground-level visibility in bean fields, coinciding with elevated nest predation likelihood. We hypothesize that the mismatch between habitat selection and fitness in bean fields was driven primarily by the speed and magnitude of crop growth in the period following initial settlement, causing declines in habitat suitability during the nest cycle. Given that rapid vegetative growth is a feature of many arable crop strains (Wilson, Whittingham & Bradbury, 2005), we suspect that similar growth-related crop changes could cause maladaptive settlement patterns in other ground-nesting bird species in arable fields. Indeed, there is some evidence that rapid growth of cereal crops can cause nest abandonment in both lapwings Vanellus vanellus and stone curlews Burhinus oedicnemus in arable farmland (Galbraith, 1989; Wilson et al., 2005), although the degree of preference for unproductive habitats in these cases is apparently low (Newton, 2004).

Within autumn-sown wheat crops, yellow wagtails tended to place nests adjacent to tramlines, where predation likelihood was highest. Donald et al. (2002) report a similar relationship in skylarks Alauda arvensis, suggesting that tramlines allow easy ground access within the otherwise dense wheat canopy (Wilson et al., 1997; Donald et al., 2001). As tramlines are used by foraging mammalian predators (pers. obs.), it is unsurprising that the nest predation rate increases with tramline proximity. Yellow wagtail pairs nesting in the crop interior (>75 cm from tramlines) achieved high rates of nest success, with no recorded cases of nest predation (of n=15 nests). Despite this, the crop interior was generally avoided for nesting, suggesting that it might represent an undervalued high-quality resource (Gilroy & Sutherland, 2007). Future studies could examine this possibility experimentally by creating linear openings within the crop interior, mimicking the appearance of tramlines without enhancing access for terrestrial predators. If successful, such a treatment could be used as a template for conservation measures aimed at increasing the productivity of declining yellow wagtail breeding populations (Grice et al., 2004; Gilroy & Sutherland, 2007). Similar targeted within-field measures have proved a successful and cost-effective means of increasing breeding productivity in skylarks within UK farmland (Morris et al., 2004).

At the population scale, impacts of maladaptive habitat selection depend not only on the degree of preference for poor quality habitats and resultant fitness costs, but also on the overall proportion of individuals settling in suboptimal habitat (Kokko & Sutherland, 2001; Kristan III, 2003). The proportionate cover of habitat therefore plays a key role in determining the population outcomes of maladaptive habitat selection. In our study area, the relative cover of bean crops was fairly small (see Table 1), such that bean fields housed a relatively small proportion of the total population. In landscapes with greater proportionate cover of field beans, the population-scale impact of maladaptive settlement could conceivably be more dramatic. However, local patterns of relative habitat preference may be poorly representative of the population as a whole (Whittingham, Wilson & Donald, 2003), and further studies would be required to determine whether the patterns observed in our study area are consistent at the population scale.

Both of the maladaptive patterns identified in this study involve habitat features that are ubiquitous in modern arable landscapes within the UK, but were effectively non-existent before the 20th century agricultural revolution (Wilson et al., 2005). These novel features represent a tiny fraction of the vast array of changes occurring in agricultural landscapes over this period (Robinson & Sutherland, 2002). It is therefore surprising that so few examples of maladaptive habitat selection have been documented among farmland species. We suspect that a more widespread awareness of this issue will lead to the discovery of additional examples, potentially assisting in the conservation of declining species. We encourage researchers to consider the use of density-based preference indices when attempting to assess relative habitat preference, provided that adequate steps are taken to exclude the confounding influence of factors such as competitive exclusion. In our study, foraging observations provided a simple but effective means of determining home range overlap, increasing confidence in density-based habitat preference assessments. Such methods are likely to broaden the capacity for researchers to detect patterns of maladaptive habitat preference in situations where individual-based preference measures are impractical.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Nicholas Watts, Sarah Nelson, Michal Maniakowski, Ute Bradter and David Gilroy for fieldwork, and Jenny Gill and Ian Newton for helpful suggestions, and all land owners that allowed access to their farms. The work was funded by Natural England and the Royal Society for the Protection for Birds through the Action for Birds in England Partnership, and also by the British Trust for Ornithology and Anglian Water. William J. Sutherland is funded by Arcadia. All methods used comply with the current laws of the country in which they were performed. The authors declare that they have no conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Aebischer, N.J. (1999). Multi-way comparisons and generalized linear models of nest success: extensions of the Mayfield method. Bird Study 46, 2231.
  • Arlt, D. & Pärt, T. (2007). Non-ideal breeding habitat selection: a mismatch between habitat and fitness. Ecology 88, 792801.
  • Battin, J. (2004). When good animals love bad habitats: ecological traps and the conservation of animal populations. Conserv. Biol. 18, 14821491.
  • Brown, J.L. & Orians, G.H. (1970). Spacing patterns in mobile animals. Annu. Rev. Ecol. Syst.1, 239262.
  • Burnham, K.P. & Anderson, D.R. (2002). Model selection and multimodel inference: a practical information–theoretic approach. 2nd edn. New York: Springer.
  • Cody, M.L. (1981). Habitat selection in birds: the roles of vegetation structure, competitors, and productivity. BioScience 31, 107113.
  • Crawley, M.J. (1993). GLIM for ecologists. Oxford: Blackwell Scientific Publications.
  • Donald, P.F., Evans, A.D., Buckingham, D.L., Muirhead, L.B. & Wilson, J.D. (2001). Factors affecting the territory distribution of Skylarks Alauda arvensis breeding on lowland farmland. Bird Study48, 271278.
  • Donald, P.F., Evans, A.D., Muirhead, L.B., Buckingham, D.L., Kirby, W.B. & Schmitt, S.I.A. (2002). Survival rates, causes of failure and productivity of Skylark Alauda arvensis nests on lowland farmland. Ibis144, 652664.
  • Eaton, M.A., Austin, G.E., Banks, A.N., Conway, G., Douse, A., Grice, P.V., Hearn, R., Hilton, G., Hoccum, D., Musgrove, A.J., Noble, D.G., Ratcliffe, N., Rehfisch, M.M., Worden, J. & Wotton, S. (2007). The state of the UK's birds 2006. RSPB, BTO, WWT, CCW, EHS, NE and SNH, Sandy, Bedfordshire.
  • Galbraith, H. (1989). Arrival and habitat use by lapwings Vanellus vanellus in the early breeding season. Ibis131, 377388.
  • Gates, J.E. & Gysel, L.W. (1978). Avian nest dispersion and fledging success in field-forest ecotones. Ecology59, 871883.
  • Gilroy, J.J., Anderson, G.Q.A., Grice, P.V., Vickery, J.A. & Sutherland, W.J. (2010a). Mid-season shifts in the breeding distribution and habitat associations of yellow wagtails Motacilla flava breeding in arable farmland. Ibis152, 90104.
  • Gilroy, J.J., Anderson, G.Q.A., Vickery, J.A., Grice, P.V., Watts, P.N., Bray, I. & Sutherland, W.J. (2008). Could soil degradation contribute to farmland bird declines? Links between soil penetrability and the abundance of yellow wagtails in UK arable fields. Biol. Conserv.141, 31163126.
  • Gilroy, J.J., Anderson, G.Q.A., Vickery, J.A., Grice, P.V., Watts, P.N. & Sutherland, W.J. (2010b). Foraging habitat selection, diet and nestling condition in yellow wagtails Motacilla flava breeding on arable farmland. Bird Study56, 221232.
  • Gilroy, J.J. & Sutherland, W.J. (2007). Beyond ecological traps: perceptual errors and undervalued resources. Trends Ecol. Evol. 22, 351356.
  • Grice, P., Evans, A., Osmond, J. & Brand-Hardy, R. (2004). Science into policy: the role of research in the development of a recovery plan for farmland birds in England. Ibis 146, S239S249.
  • Hernandez, K.M., Reece, B.A. & McIntyre, N.E. (2006). Effects of anthropogenic land use on Odonata in playas of the Southern High Plains. West. N. Am. Nat.66, 273278.
  • Johnson, M.D. (2007). Measuring habitat quality: a review. Condor 109, 489504.
  • Kaufmann, J.H. (1983). On the definitions and functions of dominance and territoriality. Biol. Rev.58, 120.
  • Kokko, H. & Sutherland, W.J. (2001). Ecological traps in changing environments: ecological and evolutionary consequences of a behaviourally mediated Allee effect. Evol. Ecol. Res. 3, 537551.
  • Kristan, W.B., (2003). The role of habitat selection behavior in population dynamics: source-sink systems and ecological traps. Oikos 103, 457468.
  • Maher, C.R. & Lott, D.F. (2000). A review of ecological determinants of territoriality within vertebrate species. Am. Midl. Nat.143, 129.
  • Mänd, R., Tilgar, V., Lõhmus, A. & Leivits, A. (2005). Providing nest boxes for hole-nesting birds – does habitat matter?Biodivers. Conserv.14, 18231840.
  • Manly, B.F.J, Miller, P. & Cook, L.M. (1972). Analysis of a selective predation experiment.Am. Nat.106, 719736.
  • Manzer, D.L. & Hannon, S.J. (2005). Relating grouse nest success and corvid density to habitat: a multi-scale approach. J. Wildl. Mgmt. 69, 110123.
  • Marchant, J. (1983). Common bird census instructions. Tring: British Trust for Ornithology.
  • Mayfield, H.F. (1975). Suggestions for calculating nest success. Wilson Bull.87, 456466.
  • Morris, A.J. & Gilroy, J.J. (2008). Close to the edge: predation risks for two declining farmland passerines. Ibis 150, 168177.
  • Morris, A.J., Holland, J.M., Smith, B. & Jones, N.E. (2004). Sustainable Arable Farming for an Improved Environment (SAFFIE): managing winter wheat sward structure for Skylarks Alauda arvensis. Ibis 146, S155S162.
  • Newton, I. (2004). The recent declines of farmland bird populations in Britain: an appraisal of causal factors and conservation actions. Ibis 146, 579600.
  • Pärt, T., Arlt, D. & Villard, M.A. (2007). Empirical evidence for ecological traps: a two step model focussing on individual decisions. J. Ornithol. 148, S327S332.
  • Pelicice, F.M. & Agostinho, A.A. (2008). Fish-passage facilities as ecological traps in large Neotropical rivers. Conserv. Biol.22, 180188.
  • Purcell, K.L. & Verner, J. (1998). Density and reproductive success of California Towhee. Conserv. Biol.12, 442450.
  • Railsback, S.F., Stauffer, H.B. & Harvey, B.C. (2003). What can habitat preference models tell us? Tests using a virtual trout population. Ecol. Appl.13, 15801594.
  • Remes, V. (2003). Effects of exotic habitat on nesting success, territory density and settlement patterns in the Blackcap (Sylvia atricapilla). Conserv. Biol.17, 11271133.
  • Robertson, G.A. & Hutto, R.L. (2006). A framework for understanding ecological traps and an evaluation of existing evidence. Ecology87, 10751085.
  • Robinson, R.A. & Sutherland, W.J. (2002). Changes in arable farming and biodiversity in Great Britain. J. Appl. Ecol.39, 157176.
  • Rodenhouse, N.L. & Best, L.B. (1983). Breeding ecology of Vesper Sparrow in corn and soybean fields. Am. Midl. Nat.110, 265275.
  • Roselaar, C.S. (1988). Yellow wagtail Motacilla flava. In Handbook of the birds of Europe, the Middle East and North Africa. Vol. V: 413433. Cramp, S., et al. (Eds). London: Oxford University Press.
  • Schlaepfer, M.A., Runge, M.C. & Sherman, P.W. (2002). Ecological and evolutionary traps. Trends Ecol. Evol.17, 474480.
  • Shochat, E., Patten, M.A., Reinking, D.L., Wolfe, D.H., Sherrod, S.K. & Morris, D.W. (2005). Ecological traps in isodars: effects of tallgrass prairie management on bird nest success. Oikos 111, 159169.
  • Smith, S. (1950). The yellow wagtail. London: Collins.
  • Sutherland, W.J. & Green, R.E. (2004). Habitat assessment. In Bird ecology and conservation: a handbook of techniques: 186202. Sutherland, W J., et al. (Eds). London: Oxford University Press.
  • Van Horne, B. (1983). Density as a misleading indicator of habitat quality. J. Wildl. Mgmt.47, 893901.
  • Whittingham, M.J., Wilson, J.D. & Donald, P.F. (2003). Do habitat association models have any generality? Predicting skylark Alauda arvensis abundance in different regions of southern England. Ecography26, 521531.
  • Wilson, J.D., Evans, J., Browne, S.J. & King, J.R. (1997). Territory distribution and breeding success of skylarks Alauda arvensis on organic and intensive farmland in southern England. J. Appl. Ecol.34, 14621478.
  • Wilson, J.D. & Vickery, J.L. (2004). Decline in yellow wagtail Motacilla flava flavissima breeding on lowland wet grassland in England and Wales between 1982 and 2002. Bird Study52, 8892.
  • Wilson, J.D., Whittingham, M.J. & Bradbury, R.B. (2005). The management of crop structure: a general approach to reversing the impacts of agricultural intensification on birds? Ibis 147, 453463.