The population declines of waders in Europe are widely considered to have resulted from habitat loss and degradation due to agricultural changes. However, recent empirical evidence suggests that levels of predation on wader nests are unsustainably high in many cases, even in some situations where breeding habitat is otherwise favourable. We review the published and ‘grey’ literature on nest predation on waders in Europe and quantify the relative importance of the major predators. Nest cameras offer the least biased method of identifying and quantifying nest predators. A small number of camera studies, in combination with others utilizing nest temperature loggers, indicate that nocturnal/mammalian predators make the largest contribution to wader nest predation. More than half of site-years or studies reviewed reported clutch failure rates of over 50% attributable to predation alone, a rate that is likely to be associated with declining populations, although parameters such as chick and adult survival will also affect population trends. Correlates of wader nest predation are documented, with time of season, field type and management, distance to habitat/field edge, wader nest density, and abundance of mammalian predators being most consistently identified. Future directions of research into wader productivity are discussed, and we suggest that studies quantify additional life-history parameters such as chick survival, as well as examining the predator community, wherever possible.
These declines in wader populations have been attributed to alterations of habitat associated with agricultural intensification, notably drainage and the use of inorganic fertilizers, that have rendered habitat less suitable (Shrubb 1990, Donald et al. 2001, Wilson et al. 2004). There has been an emphasis on the restoration and creation of habitat, such as extensively managed lowland wet grassland, in an effort to increase populations or to stem their declines, and there is mixed evidence that these are effective in increasing wader populations (Verhulst et al. 2007, Wilson et al. 2007).
It is undeniable that appropriate habitat management is necessary to maintain or increase populations of waders; however, increased nest predation has been suggested as a possible additional cause of wader population declines (Bellebaum 2002b, Chamberlain & Crick 2003, Milsom 2005). Predators may cause prey populations to decline to extinction, or to stabilize at lower levels, where they may be more susceptible to extinction from other causes. Alternatively, populations that have declined due to other causes may be more likely to suffer increased rates of predation. There are several possible mechanisms whereby agricultural intensification may itself lead to increased rates of predation on wader nests: smaller and more dispersed wader populations may be less effective at deterring predators by mobbing; anti-predator vigilance may be compromised in rapidly growing, taller swards; and nests may be more vulnerable to predation due to reduced crypsis in homogeneous swards produced by high fertilizer inputs (Shrubb 1990, Evans 2004, Newton 2004, Whittingham & Evans 2004, Wilson et al. 2005b). Additionally, numbers of some nest predators may have increased in response to agricultural intensification.
A further issue related to nest predation is identification of nest predators. Information on the identity of nest predators is vital to inform management intended to decrease levels of nest predation. Nest predation is not easily observed but the means of identifying nest predators and of quantifying the contribution of the various predators of wader nests have improved in recent years (Sabine et al. 2005, Bolton et al. 2007a, Pierce & Pobprasert 2007).
We first document the means by which nest predators have been identified and quantified, and comment on the reliability of these means. We then review the evidence that nest predation is a major issue for waders in Europe, and identify factors that have been shown to affect rates of nest predation.
Using the Web of Science database and the online search engine Google Scholar, peer-reviewed papers on predation of wader nests and the identity of predators were located. Tables of Contents of relevant journals and the catalogue of the RSPB library were searched for relevant articles and reports. Literature cited in the papers and reports identified by our primary search methods were examined. Experts in the field of nest predation were consulted directly to locate other literature and unpublished data that may not have been located during the online and database searches. The review was restricted to waders nesting in Europe (excluding the high Arctic) about which there exists a relatively large amount of information, and included studies published up to 2007. Studies from outside Europe are referred to occasionally, where their inclusion helps to clarify some of the concepts discussed, particularly with regard to the identification of nest predators.
Studies were included in the review if the Mayfield, or exposure days, method of determining daily predation rate (DPR) or daily failure rate (DFR) had been followed (Mayfield 1961, 1975). Studies presenting DFR were included if DPR could be calculated from the data presented or if DFR was presented and predation was stated to be the major cause of nest failure (studies where this was the case are identified in Appendix 2). A very small number of studies that reported DFR but for which the contribution of predation could not be determined were excluded. The exposure days method uses the number of failed/predated nests divided by the number of days that nests are under observation to calculate DFR or DPR, which is then used to calculate hatching rates by raising 1 – DFR or 1 – DPR to the power of the number of days of laying/incubation (Mayfield 1961, 1975).
The means of presenting results varied, and not all studies examined listed DPR (or DFR), or daily survival rate (1 – DPR or 1 – DFR). In some cases DPR was back-calculated from percentage hatching rates using the number of laying/incubation days presented in the study, or read off graphs of DPR/DFR. Studies quantifying the contribution of nest predators were included in the review of the identity of nest predators, even if the study did not use the exposure days method to calculate DPR. To make valid comparisons among studies, DPR/DFR was converted into hatching success (based solely on losses to predation, where these were separated) for species by using standardized risk days for each species. For the laying period, this was 1 day per egg in the modal clutch size, and for the incubating period, the mid-point of the incubation period reported in Snow and Perrins (1998).
Studies of predation of artificial nests were not included in this review, neither to identify nest predators nor to report DPR, as there are serious problems in applying the results of artificial nest studies to real nests. Moore and Robinson (2004) concluded that predation rates on real and artificial nests differ in unpredictable and inconsistent directions, primarily because each type attracted different predators. This finding has been borne out in several studies of predation of artificial wader nests (Berg 1996, Valkama et al. 1999, Ottvall 2005b).
IDENTIFICATION OF PREDATORS OF WADER NESTS
Simple observations of nest predation are likely to be biased towards diurnal predators, which tend to be avian. The response of nesting waders (e.g. mobbing, fleeing or feigning incubation) to other species has also been used to infer predation risk (Berg et al. 1992, Sasvári & Hegyi 2000, Šálek & Cepáková 2006). However, this approach suffers from the same potential bias as observations of nest predation. Systematic nocturnal observations of the responses of nesting waders to predators have been made in one study; Northern Lapwings Vanellus vanellus responded vigorously to the presence of Red Foxes Vulpes vulpes (Seymour 1999). Methods of quantifying predator contributions include dietary studies of predators, the examination of nest remains, recording the timing of predation events and the use of nest cameras (Appendix 1).
Evidence of egg predation by examination of predator diet is difficult to obtain, as egg remains are infrequent in digestive tracts and scats. It has been stated that egg shells are generally dissolved in fox stomachs within 4 h (Lever 1959), although remains have been identified in scats (Baker et al. 2006), and another source states that eggshells are not usually dissolved by stomach acids (Neal & Cheeseman 1996). In fact, the lack of egg remains in guts and faeces may reflect their lack of importance in the diet. A review of the importance of birds (by weight) in the diet of various predators in the British Isles did not distinguish eggs as a separate category (Cotgreave 1995). Data on the importance of eggs in the diet of potential nest predators indicate that for all species except the Stoat Mustela erminea, birds’ eggs form a very small proportion of the diet (Lockie 1956, Yom-Tov 1975, Tapper 1976, Shepherdson et al. 1990, McDonald et al. 2000, Baker & Harris 2003, Hounsome & Delahay 2005). There are possible biases in dietary assessment due to differential passage to stomach and faeces, and because stomach contents may come from a biased sample of the population (Cavallini & Volpi 1995). In addition, large numbers of samples are required, which are likely to come from a wide range of habitats, so they are limited in terms of identifying prey items that are important at smaller scales. Thus, dietary examination is not a suitable means of quantifying the contribution of nest predators.
Nest and eggshell remains have been widely used as a means of identifying nest predators (e.g. Green et al. 1987, Valkama & Currie 1999, Whittingham et al. 2002, Bellebaum & Boschert 2003), as the patterns of egg and nest damage tend to vary among classes of predator. In some cases the identity of mammalian predators can be deduced from the distances between incisor toothmarks. Such methods have shown that mustelids are important ground-nest predators in German wet grassland; their importance was greater in regularly flooded areas, where they tended to replace Foxes (Bellebaum 2002a). However, caution must be exercised when inferring the relative contribution of different predator species from such evidence, as some predators are less likely to leave signs than others, distinguishing between predators is often imprecise and in most studies no signs are left at a large proportion of nests (Lariviere 1999). Comparability between studies using nest remains is difficult: for example, one study categorized all of 416 Avocet Recurvirostra avosetta nest predation events as either mammalian or avian (Lengyel 2006), while other studies have reported varying proportions of unknown nest predators. This may honestly reflect the evidence present at the nests in the various studies, or may indicate that various researchers accept different levels of evidence to categorize nest predators. The use of nest cameras (described below) should allow calibration of the use of nest remains as identifiers of predators.
Wax eggs have been used to identify predators of nests of some bird species (e.g. Anthony et al. 2006), but have generally been used in artificial nests. They have been used in real wader nests in one study where nest predation at one site following addition of wax eggs appeared to be heightened (Grant et al. 1999), and another where all wax eggs disappeared from predated nests: this was taken as an indication that mammals were the predators, as birds were not thought to be strong enough to remove the peg attached to the egg (Valkama & Currie 1999).
Timing of predation (temperature loggers)
Nest temperature loggers have sensors that are placed within nests and record nest temperature continuously at short intervals. These data can provide information on the time of day at which nests were predated. Nest predation events are generally classified as nocturnal or diurnal (although some studies also distinguished predation during twilight); nocturnal predation events can be fairly reliably ascribed to mammalian species, but although diurnal predation is more likely to be avian, mammals also take eggs during daylight (Boschert 2005).
Remotely operated nest cameras provide direct, and therefore the most reliable, information on the identity of nest predators. The earliest photographic systems were based on film cameras triggered by activity, such as removal of eggs, at the nest (Danielson et al. 1996, Sawin et al. 2003). More recently, continuous time-lapse recording has been widely employed in avian studies, although such systems are large and heavy and require frequent maintenance (Pietz & Granfors 2000, Renfrew & Ribic 2003). The advent of miniature cameras employing digital technology and sophisticated software-based triggering systems such as Video Motion Detection have led to the development of more compact low-maintainence systems (Bolton et al. 2007a).
The presence of temperature loggers and nest cameras, and the attendant disturbance in placing these devices, might be expected to increase rates of predation. However, this was not found to be the case for temperature loggers in Piping Plover Charadrius melodus and Long-billed Curlew Numenius americanus nests in the USA (Hartman & Oring 2006, Schneider & McWilliams 2007), for video cameras at wader nests in New Zealand (Sanders & Maloney 2002), or for digital nest cameras at Lapwing nests in the United Kingdom (Bolton et al. 2007a).
A final way of identifying predators is to document changes in populations and/or breeding success following changes in predator abundance (generally achieved through control programmes). In the UK, decreases in Lapwing nest predation followed control of Foxes and Carrion Crows Corvus corone in wet grassland areas (Bolton et al. 2007b); however, as this study involved control of both mammalian and avian predators, it does not establish which were most important. In North America, exclusion solely of mammalian predators led to a significant increase in nest success in Piping Plovers, while exclusion of avian predators saw no further increase in nest success (Ivan & Murphy 2005). In Denmark, exclusion of mammalian predators increased Lapwing hatching success, although continued nocturnal predation suggested that the exclusion had been only partly successful (Olsen 2002). Given the range of potential predators, and the impracticality of designing an experiment to control each in turn, predator removal is unlikely to be a good way of identifying nest predators, and would more profitably be pursued as a management tool once predators have been identified using other methods. This has occurred in Scotland's Western Isles, where Hedgehogs Erinaceus europaeus were identified as major nest predators from nest remains and nest temperature loggers, and their exclusion led to increased wader nest success (Jackson 2001).
IDENTITY OF WADER NEST PREDATORS
The three major methods of identification and quantification of wader nest predators offer different levels of precision (Fig. 1). Across all studies, the use of nest remains resulted in over 45% of predation events being classified as unknown. The predator was identified to species in only 12.1% of predation events. Birds identified to species were Common Gull Larus canus (n = 70), Yellow-legged Gull L. michahellis (n = 38) and Carrion Crow (n = 1), each species being identified in a single study (Robson 1998, Rusticali et al. 1999, Jackson 2001). Mammal species identified were Hedgehog (n = 75), Fox (n = 58), Stoat (n = 23), Mink Mustela vison (n = 3) and Polecat M. putorius (n = 2). The high number for Hedgehog comes from a single study where the mammalian predator community was very limited (Jackson 2001). Of all predation events where the predator was at least classified (n = 1067), mammals accounted for 70%.
Temperature loggers and nest cameras have a much lower proportion of unknown classifications of nest predators as compared with examination of nest remains, although temperature loggers cannot identify predator species (Fig. 1). Over all studies, nocturnal predation predominated, even in situations where the expected predators were avian (e.g. Olsen 2002). In Dutch meadows and grassland, nocturnal predation of meadow bird nests was most important where predation rates were high (> 50%), and where predation rates were lower, nocturnal and diurnal predation contributed equally (Teunissen et al. 2005). Temperature loggers in combination with nest remains were used to identify Hedgehogs as the major predator of wader nests in machair on the Scottish island of Uist (Jackson & Green 2000). Of 216 nest predation events identified using nest cameras, predominantly in wet grassland, 132 were by Foxes, 21 by Stoats, 15 by Crows, 12 by Badgers Meles meles, eight by Hedgehogs, small numbers by other predators, and in 16 cases the predator could not be determined (Blühdorn 2002, Teunissen et al. 2005, Sharpe 2006, Bolton et al. 2007a, RSPB unpubl. data).
Predators of nests may vary greatly between sites, even where habitat and management appear similar (Grant et al. 1999). However, although the range of predators is great, there is growing evidence from remote monitoring devices, where bias is minimized, that in many situations the majority of predation occurs at night and is therefore attributable to mammalian species. The use of temperature loggers initially, and nest cameras more recently, has shown that the widely held belief that birds (particularly corvids) are the major predators of wader nests is frequently not true. The limited quantitative information currently available suggests that the most important nest predator species include Fox, Badger and Stoat.
RATES OF PREDATION ON WADER NESTS
Fifty-seven papers, theses and reports were identified that documented DPR, or failure rates where predation was the major cause of failure, in 17 wader species. We were also granted access to further unpublished data (Appendix 2). Some studies distinguished between DPR during laying and DPR during incubation, generally finding the latter to be lower (Beintema & Müskens 1987, Grant et al. 1999, Verboven et al. 2001, Pearce-Higgins & Yalden 2003, Cuervo 2005), although one study found no difference (Berg 1992), and another found that predation rates were lower during laying (Sharpe 2006). Separate DPR values for laying and incubating were used, when presented, to calculate hatching success. In some studies, DPR values were presented for site-years, whereas other studies presented DPR for the overall study, or amalgamated DPR values, for example by habitat or by management (see Appendix 2 for the number of reported values from each study). Considerable variation in hatching success has been reported for wader species across a range of habitats throughout Europe; more than 50% of nests were predated in 55.3% of site-years or studies reviewed (n = 544). For individual species with at least 10 reported values these percentages were: Black-tailed Godwit Limosa limosa, 39.7% (n = 73); Northern Lapwing, 57.7% (n = 305); Redshank Tringa totanus, 60.1% (n = 46); Avocet, 22.2% (n = 18); Eurasian Curlew Numenius arquata, 71.4% (n = 21); Dunlin Calidris alpina, 63.6% (n = 11); Oystercatcher Haematopus ostralegus, 56.3% (n = 32) and Ringed Plover Charadrius hiaticula, 90% (n = 10).
IMPACT OF NEST PREDATION ON WADER POPULATIONS
Wader declines are generally believed to result from poor breeding success rather than low annual survival rates of adults or immatures (Peach et al. 1994, Catchpole et al. 1999, Hötker et al. 2007, but see Johansson 2001). It can be difficult to ascribe changes in populations definitively to changes in nest predation, as other demographic parameters can have a large influence. Most waders will re-nest, either as multiple brooding or in response to nest failure, which may compensate, to some degree, for low nest survival (Beintema & Müskens 1987). Chick survival is an important demographic parameter affecting productivity that is more rarely measured than nest survival. Despite high levels of nest predation, chick predation was considered to have had more impact on wader productivity in meadows in the Netherlands (Teunissen et al. 2005), and productivity of Avocets in the Wadden Sea coast of Germany was not related to hatching success, but was positively related to chick survival (Hötker & Segebade 2000). Adult survival rates can have a large effect on the productivity required to maintain a stable population (Ottvall & Härdling 2005). Population trends are frequently heavily affected by emigration and immigration, so changes in population size may not reflect productivity on the local scale (Bellebaum 2001). The lack of a relationship between DPR and population trends is therefore not surprising, as in the case of Lapwings across upland sites in Great Britain (O’Brien 2001). Nevertheless, numerous studies have drawn attention to high rates of nest predation as responsible, at least in part, for low productivity in waders (e.g. Galbraith 1988, Green 1988, Baines 1990, Jönsson 1991, Grant et al. 1999, Hart et al. 2002, Ottvall 2005b, Smart 2005, Thyen & Exo 2005, Junker et al. 2006). The Temminck's Stint Calidris temminckii population of Bothnian Finland declined by nearly 40% between 1987–95 and 1999–2002 (Rönkäet al. 2006). This decline coincided with a significant increase in DPR, which increased in both absolute magnitude and proportion of nest failures from the periods 1983–91 to 1992–2001. From reported adult and juvenile survival rates, it was estimated that nest survival would need to exceed 49% to maintain the population; this figure was not reached due to predation, without considering losses from other causes (Rönkäet al. 2006).
The impact of DPR on annual wader productivity, and hence their population trends, depends to a large extent on the other components of annual breeding success and annual survival, as described above. These parameters are relatively well studied for Lapwings, and we used a simple deterministic model to examine the impact of varying levels of nest survival on annual reproductive output (the number of chicks fledged per pair) under different levels of chick survival. The model assumed that 69% of females re-laid following nest failure (Baines 1989, Hegyi 1996; Hegyi & Sasvári 1998), and would produce a maximum of three clutches in a breeding season, but that loss of a brood was not followed by re-nesting (Shrubb 2007). The average number of hatched eggs among successful clutches (i.e. accounting for non-viable eggs) was set at 3.63 (M. Bolton unpubl. data). The model considered nest survival rates (from laying to hatching) and chick survival rates (from hatching to fledging) that reflected the range of values in the published literature, namely 0–100% and 0–50%, respectively. The annual reproductive output per female under each combination of nest and chick survival rates is shown in Figure 2. The shaded contour indicates the level of productivity that is required to maintain a stable population (0.6–0.8 chicks fledged per female), given published estimates of adult and first-year survival rates (Peach et al. 1994, Catchpole et al. 1999). The curvilinear relationship between nest survival and annual productivity contrasts with the linear relationship between chick survival and productivity, and reflects the effect of re-nesting following nest loss. This simple model suggests that for chick survival rates of 25%, rates of nest loss (to all causes combined) up to 50% are sustainable, but in situations of lower chick survival, higher rates of nest survival must be achieved to maintain stable Lapwing populations.
FACTORS AFFECTING RATES OF NEST PREDATION
Nest predation rates might be expected to vary according to habitat variables, predator community and behaviour of the nesting waders. Many studies have attempted to relate at least one of these elements to DPR (Table 1). Some studies have also related these habitat variables to the distribution of wader nests; however, factors that affect nest distribution do not necessarily affect DPR (Berg 1992, Grant et al. 1999, Thyen & Exo 2005). This may be because waders make decisions on nest-sites based on factors other than minimizing nest predation, or because waders make decisions based on a perceived level of predation risk that is not realized. Selection of non-ideal habitat for nesting has been observed in waders (Székely 1992, Kleijn et al. 2001), where waders have selected areas with more food for adults, although reproductive rates were lower, in the latter case due to high predator densities.
Table 1. Number of studies finding effects of variables on wader nest predation rates.
Effect not found
Includes one study where the effect varied significantly in different directions in different years.
Studies tended to investigate distance to more than one feature; if one of these was significant it is included here.
Where studies investigated the effects of multiple predators, if one was significant it is included here.
The time of season at which nests are present might be expected to relate to DPR, as predator abundance and/or behaviour, or nest susceptibility to predation changes (e.g. through changes to vegetation) during the course of the nesting season. Studies have considered this in different ways: as a continuous variable, comparing first and replacement clutches, or as several categories. For four wader species in wet grassland in the Netherlands (Common Snipe Gallinago gallinago, Black-tailed Godwit, Lapwing and Redshank), DPR was highest at the beginning and the end of the season (Beintema & Müskens 1987). The authors suggested that nest susceptibility due to changes in crypsis were unlikely to explain the peak in nest survival in the middle of the season, but that a swamping effect may have been acting, or that DPR depended more on the predators themselves. A similar peak in nest survival in the middle of the season was observed in coastal meadows in Denmark for Lapwing and Dunlin (Thorup 1998). At the same site, nest failure was higher for early nests of Ruff Philomachus pugnax, Black-tailed Godwit and Snipe, a trend also observed in Snipe in wet grassland (Green 1988) and Ringed Plovers in coastal dunes and beaches (Pienkowski 1984). Elsewhere, DPR increased with the progression of the season for Redshank and other species on coastal meadows and saltmarsh (Ottvall 2005a, Ottvall et al. 2005, Thyen & Exo 2005), but did not do so for a range of other species in various habitats (Galbraith 1988, Berg et al. 1992, Grant et al. 1999, Pearce-Higgins & Yalden 2003, Wallander & Andersson 2003), and was higher both earlier and later in the season, depending on the year, in one study (Blühdorn 2002). First clutches were less likely to fail for Redshank and Ringed Plover in machair in the Western Isles of Scotland, but three other species at the same sites did not show the same pattern (Jackson & Green 2000). Although several studies have found that DPR is temporally variable, the underlying mechanisms are not known. Examination of interactions between time and other factors may help to determine these mechanisms.
Nest predation rates may vary among habitats such as wet grassland, upland moorland and arable fields. Some studies presented habitat-specific DPR values. For others, habitat classification was assigned based on the site descriptions provided. The distribution of hatching success rates shows that arable habitats have a higher proportion of studies reporting high hatching success, and that upland habitats have the lowest proportion (Fig. 3). Most of the site-years and studies categorized as upland habitats are rough grazing or marginal farmland, rather than moorland.
The tendency for grass fields to suffer higher levels of nest predation than arable fields has also been documented several times by studies that have analysed predation rates separately. However, although arable fields typically have lower DPR than grass fields (Galbraith 1988, Baines 1990, O’Brien 2001, Seymour et al. 2003, Sharpe 2006, Sheldon et al. 2007), this is not universally the case (Berg et al. 1992, J. Bellebaum & C. Bock unpubl. data), and one study found higher DPR in arable than in wet grassland (Ratcliffe et al. 2005, S. Schmitt unpubl. data). The scale of study may be important; in Sweden, DPR was higher on Curlew nesting in a mixed farm landscape than in a predominantly arable landscape, although within the mixed farm landscape there was no difference in DPR between arable fields and grassland (Berg 1992). Similarly, DPR on Lapwing nests in upland Britain was lower at sites where more arable land was present and where more improved grassland was present (O’Brien 2001).
Other differences in DPR by habitat have been reported: Whimbrel Numenius phaeopus nest survival was higher in alpine heath than in forest, mire and dry heath (Pullianen & Saari 1993); Redshank DPR decreased along a saltmarsh succession gradient (Thyen & Exo 2005); and Temminck's Stint DPR was higher in artificial than natural habitats in one time period examined (Rönkäet al. 2006). The effect of topography on DPR was examined by one study in British moorland, where Golden Plovers Pluvialis apricaria selected flat areas to nest and nests survived longer on flat areas than on slopes (Whittingham et al. 2002). This may be because nests on flat ground may afford a greater field of view for incubating birds, improving their ability to detect predators.
Field management can also affect DPR. Lapwing nest DPR in upland Britain was higher in improved pasture than in unimproved pastures and meadows (Baines 1990), and in another study, Lapwing nests on pastures suffered higher DPR than those on meadows (Fletcher et al. 2005). Lapwing nest DPR on organic and conventionally managed arable fields in the Netherlands did not differ (Kragten & de Snoo 2007). Grazed coastal marshes, or marshes grazed at higher intensity, reported higher DPR for Oystercatcher and Lapwing (Bruns et al. 2001, Hart et al. 2002). This may be because disturbance of incubating adults may help predators to locate clutches, although the effect was not observed for several wader species on coastal meadows (Ottvall 2005a) or for Lapwing in upland grassland (Fletcher et al. 2005).
Differences in wader nest DPR between habitats are unsurprising, and presumably arise from the different predator communities present. Variation between field types within the same study sites probably reflects the behaviour of predators: for example, arable fields are less attractive foraging areas. There may be some potential to reduce nest predation in some grassland areas by lowering grazing intensity, although the evidence for this effect is not conclusive.
Nest concealment has been recorded in several studies, as it might be expected to be important in determining DPR. However, although it appears to affect nest placement, only one study found that DPR (of Lapwing nests in upland habitats) was higher where vegetation was shorter (O’Brien 2001). All other studies have found no effect of nest crypsis, although the means of measuring this vary (Green 1988, Grant et al. 1999, Whittingham et al. 2002, Ottvall 2005a, Ottvall et al. 2005, Thyen & Exo 2005, Büttger et al. 2006, Smart et al. 2006, Cepakova et al. 2007, MacDonald & Bolton 2008). The fact that nest concealment affects nest distribution suggests that waders are attempting to minimize nest predation by predators foraging using sight, i.e. birds. The lack of relationships between nest concealment and DPR suggests that this is ineffective, that the predators are largely using non-visual methods or that nest predation is incidental during other foraging activity.
Distance to habitat features (including predator perches)
The effect of distance from various features on DPR has also been investigated by several studies. Distance to Common Kestrel Falco tinnunculus nests had no effect on Curlew nest DPR in Finland (Valkama et al. 1999). Predation on Lapwing nests in wet grassland was not related to distance to Fox earths (MacDonald & Bolton 2008, J. Bellebaum & C. Bock unpubl. data); nor was there any effect of proximity to Crow nests (MacDonald & Bolton 2008). Lapwing nests closer to Crow nests at one site in upland Britain were more likely to be predated, but a study of multiple sites found no effect (O’Brien 2001). Presence and proximity of nests of other avian predators had conflicting effects on DPR of Lapwing nests. Sites with Buzzard Buteo buteo nests suffered higher nest predation, but there was no effect of proximity to Buzzard nests, whereas there was an effect of proximity to Common Gull colonies, but no difference between sites with and without colonies. Sites with rookeries had lower DPR, but nests closer to rookeries had higher DPR (O’Brien 2001).
Higher DPR close to habitat edges suggests that predators are more likely to forage along the edges of fields or other linear features, as has been observed of Foxes in wet meadows in the UK (Seymour et al. 2003). The range of distances from edges/perches available to nesting waders will vary between sites. This may help to explain why around half of studies found this to be a significant correlate of DPR, although site-specific factors (such as the predator community or the nature of habitat edges) may mean that this factor may indeed only be important at certain sites. That relatively few studies found an effect of distance to vantage point provides further evidence that nest visibility tends to be unimportant.
It has been suggested that nesting in close proximity to other species that defend their nests aggressively is a strategy employed by some waders, notably at higher latitudes (Brearey & Hildén 1985, Larsen & Moldsvor 1992, Nguyen et al. 2006, Quinn & Ueta 2008). Curlews in Finland nested closer to Kestrel nests than would be expected by chance, possibly due to reduced nest predation risk from other predators (Norrdahl et al. 1995). Combined nest density of four wader species (Lapwing, Black-tailed Godwit, Oystercatcher and Redshank) explained variation in nest predation rates in Lapwing and Redshank in the Netherlands (Beintema & Müskens 1987). In Denmark, exclusion by Lapwings and Black-tailed Godwits of hunting Common Gulls from areas where Ruffs and Dunlins were nesting was thought to be the cause of lower DPR for their nests (Thorup 1998). However, in Sweden, non-aggressive wader species did not breed more successfully when within the territories of aggressive ones (Flodin et al. 1995).
Two possible mechanisms explain the negative relationship between nest density and DPR: waders nesting at higher density are better able to deter predators; or waders settle to nest at higher densities in areas of low predation pressure. The inconsistency in detecting a relationship between DPR and nest density may reflect the predator communities at study sites, as some predators may be more easily deterred than others.
Predator abundance and behaviour are likely to play a large part in determining the patterns of nest predation. Fox presence (but not avian predator density) predicted Avocet hatching success at sites where predation was the major cause of nest failure (Hötker & Segebade 2000). Fox density predicted Lapwing nest DPR in upland farmland in Great Britain (O’Brien 2001), and in wet grassland in Germany (J. Bellebaum & C. Bock unpubl. data), although not in Britain (MacDonald & Bolton 2008). An effect of avian predator density has been less commonly observed, although areas with hunting Common Gulls in Danish brackish meadows had higher DPR for Ruffs and Dunlins (Thorup 1998). Exclusion or reduction in numbers of predators has been observed to improve wader nest survival on machair in Scotland (Jackson 2001), and on wet grassland in Denmark (Olsen 2002) and in Britain (Bolton et al. 2007b). Nest survival of waders on islands without mammalian predators has been found to be high relative to mainland sites (Köster et al. 2001, Büttger et al. 2006). There is some evidence that populations of nest predators have increased in Europe (Panek & Bresinski 2002, Bellebaum 2003, Langgemach & Bellebaum 2005) from various causes, including changes in hunting practices and eradication of rabies.
Predator behaviour in relation to the abundance of alternative prey may also affect nest predation rates. Prey switching has been suggested as the cause of high DPR in several bird groups, notably where small mammal populations vary greatly (Angelstam et al. 1984, Summers & Underhill 1987, Bêty et al. 2002, Smith et al. 2007). In the Netherlands, annual nest predation rates of Lapwing and Black-tailed Godwit (but not of Redshank and Oystercatcher) were negatively correlated with Field Vole Microtus arvalis densities (Beintema & Müskens 1987). By contrast, increased abundance of alternative prey was suggested as a possible explanation for high DPR on Lapwing nests in German wet grassland; more abundant predators were possibly more likely to encounter wader nests while foraging for other prey (Köster & Bruns 2003).
Mammal predator presence/abundance was related to DPR in most studies where it was examined, but measurement of mammal abundance at a scale appropriate to individual nests or groups of nests is difficult, especially where multiple predators may be present. Assessment of predator behaviour relevant to wader nest survival presents further problems. Eggs are not a major food item of any of the nest predators identified, with the possible exception of Stoats, and this suggests that nest predation is incidental (Vickery et al. 1992, Jackson 2003), unless predators switch to searching for nests when alternative prey is scarce.
CONCLUSIONS AND FUTURE DIRECTIONS
Identification of wader nest predators
Nest cameras offer the least biased method of identification of wader nest predators, as in almost all cases the predator species can be identified, whereas other methods tend to categorize predators (e.g. nocturnal or diurnal) and/or have a high proportion of unidentified predators. The number of studies using nest cameras so far is small, and the nest predator community may be very different in other areas. We recommend that cameras are used where possible to identify nest predators, and that identification of nest predators should precede management to reduce nest predation.
Nest predation and wader population trends
Almost all studies of nest survival of waders have found that predation is the principal cause of nest failure. Other causes of failure are either unusual (but not necessarily insignificant), such as flooding, or more easily addressed by human management, such as destruction by agricultural activities. If low hatching success is having an impact on wader populations, predation appears to be the most widespread cause. However, the effects of nest failure on population trends are difficult to establish because other parameters, such as chick and adult survival and immigration/emigration, may also be important. Nest survival is more readily quantified in the field than other population parameters, and this is reflected in the large number of studies that have reported hatching success relative to those reporting fledging success, adult survival or overall population trends.
It is possible that nest predation is compensatory rather than additive. If chicks that would have hatched from predated clutches are doomed to death from other causes, or if fledged birds cannot find suitable breeding habitat, then reducing nest predation will not be sufficient to increase population sizes. However, the evidence presented in this review and in individual studies shows that in some situations, stable populations are not possible at current rates of nest predation, regardless of other factors. In these situations, reduced nest predation may be considered necessary, but not necessarily sufficient, to stabilize/improve population trends.
Factors affecting wader nest predation
Given that nest failure appears to be an important demographic parameter for waders in many cases, and that nest predation is usually the major cause of nest failure, variables that account for variation in DPR are of interest. From this review, several factors have been consistently reported as predictors of wader nest DPR, although none universally, which is unsurprising in a review of so many different species and habitats. Variables that were consistently reported as significant correlates of DPR were field type and presence/abundance of mammalian predators. Nest timing, grazing intensity and wader nest density were also significant correlates in more than 50% of studies that tested their relationships with DPR.
There is limited support for some of the possible mechanisms by which agricultural intensification may increase DPR. Wader populations that are reduced by agricultural intensification appear to be less effective at deterring predators; DPR tends to be higher in smaller or sparser nesting colonies. Reduced nest crypsis in homogeneous swards does not appear to increase DPR, while changes in anti-predator vigilance due to rapidly growing, taller swards have not been measured. Increases in populations of nest predators have been reported, and some studies have related predator abundances to DPR. However, establishing causality between agricultural intensification, predator populations trends and trends in nest predation rates will require further work.
There is a large literature on wader nest predation in Europe. However, we suggest a broader approach to studies of wader productivity and population trends, and we identify three main areas where we suggest that efforts should be concentrated. First, nest predation has been established as the major cause of nest failure in most cases, and it has been suggested as a major cause of insufficient productivity. However, the effect of changing DPR on overall productivity has rarely been assessed, in part because productivity is often measured using different methods than nest success (such as the proportion of adult pairs with young at the end of the breeding season). The use of Mayfield methods of estimating chick survival, which has been done in some cases (e.g. Grant et al. 1999, Pearce-Higgins & Yalden 2003, Bolton et al. 2007b), should allow simulated manipulation of DPR to determine what reductions in nest predation would sufficiently increase productivity. This may indicate whether chick survival (or even nest losses to other causes) is more important as a driver of productivity. A modelling approach of this sort (excluding sources of nest/chick loss in turn) in the Netherlands found that reducing avian predation on chicks was the most likely way to reverse a negative population trend in Bar-tailed Godwit and Lapwing (Teunissen et al. 2005). Despite the more intensive and expensive field methods required to quantify chick and adult survival, without such data the importance of each component to population trends is difficult to establish. We recommend that studies of wader breeding biology incorporate chick and adult survival, and population trends, either measured in the field or from published values, and consider the impact of varying rates of nest predation to assess its importance in affecting populations.
Secondly, if nest predation is established as having an effect on productivity or population trends, what factors predict the likelihood of nest predation? From our review, relationships between DPR and several variables, such as nest density, distance to habitat features and predator abundance, have been identified. However, at present these are almost exclusively correlative relationships, and it would be valuable to establish causal mechanisms for the relationships. Manipulative studies to uncover such mechanisms are likely to be expensive and time-consuming, but are important if we are to understand the causes of unsustainable rates of predation on wader nests where they occur. Information gained from correlative studies can be used to help design manipulative studies.
Finally, relating nest predation to habitat variables is relatively straightforward compared with establishing relationships between nest predation and the abundance and behaviour of potential nest predators. However, there is a need for an improved understanding of the nest predator community. A small number of studies that measured mammalian abundance or presence found that it was significantly related to DPR, but knowledge about what factors increase or decrease the likelihood of predation by various predators is very limited. Equally, research into predator ecology and behaviour is not generally intended to determine the importance of wader eggs in predator diets. Technological advances such as nest cameras have improved the identification of wader nest predators, and will continue to do so. This knowledge should allow targeted studies of the predators identified as important, to determine what influences their abundance and/or behaviour in areas of nesting waders. Manipulative studies, whether based on predator control, habitat manipulation or both, could then be designed to assess the efficacy of various management options; those found to be successful can be incorporated into management to increase wader productivity in appropriate situations.
This study was funded by Defra as part of a larger project carried out under contract to the Centre for Ecology and Hydrology (contract number C03043), and we are grateful to Richard Stillman, Sarah Durell, Andy West and Richard Caldow for comments on an earlier version of this review. Jochen Bellebaum and three reviewers made valuable comments on a draft of this manuscript. We are particularly grateful to those people who have provided unpublished data for this review: Jochen Bellebaum, Sabine Schmitt, Sarah Eglinton and Mark O’Brien. Bianca Duijs, Sabine Schmitt, Jochen Bellebaum and Preben Clausen helped with translations of some publications.
Table Appendix1.. Studies identifying and quantifying predators of wader nests.
Number of values reported: site-years where these were available, or overall values of multiple site/year studies, or values for different habitats, where these were presented separately.
Daily predation rate. Where more than two values were presented, the median of the values reported is presented, with the range in brackets. Values in italics are daily failure rate (with predation the major cause of failure). Values presented here are for incubation only, but where reported, values during laying were used to calculate hatching rates. Site-years with fewer than five nests are not presented.
These studies are presented twice in this table as they include two very different studies in the same publication.
Values presented include our estimates (for example, reading values off graphs, or back-calculating from hatching success).
Species were combined to determine predation rates, but hatching rates were calculated separately for each species (based on different number of exposure days).
This study presented only hatching success, and not risk days; 31 risk days were assumed.