Population changes of breeding waders on farmland in relation to agri-environment management

Abstract

Capsule Population response of breeding waders to agri-environment management varied between management options and species; implementation has been on too small a scale to reverse national population declines.

Aims To test whether numbers of five breeding wader species have shown a more positive response between 1992 and 2005, at sites with appropriate agri-environment management, than at sites that have remained outside such schemes.

Methods Using data from 60 pairs of farmland study areas in Scotland first surveyed in 1992/93, before agri-environment scheme (AES) implementation, and again in 2005, after scheme implementation, we tested at both site and field scales whether changes in the abundance of five breeding wader species were associated with AES management options designed to benefit these species.

Results Changes in breeding wader abundance were more positive on sites in AES, especially for Northern Lapwings Vanellus vanellus and Common Redshanks Tringa totanus, even though management had not been targetted specially at breeding waders on those sites. However, AES management was associated only with modest population increase for Common Redshanks, and a reduction in the magnitude of decline for Northern Lapwings. At the field scale, there was evidence for Northern Lapwings, Common Redshanks and Common Snipe Gallinago gallinago that options which limited grazing and other agricultural activity were associated with more positive outcomes than those which also manipulated water levels.

Conclusions AES management for breeding waders slowed, and in some cases reversed, breeding wader decline at field and farm scales. These benefits were from options that limited grazing and agricultural operations during the breeding season, but not those that also aimed to raise water levels. A possible explanation is that when wetland options are applied to agriculturally marginal fields, grazing reduction or abandonment, and succession to rank vegetation cover then occur over the course of 5-year agreements, with detrimental effects for breeding waders. Verification arrangements need to be robust enough to guard against this. Levels of agri-environment provision in 2005 were too limited and too poorly targeted at remaining key areas for breeding waders to be able to halt or reverse national population declines.

By 2006 almost four billion euros were paid annually through agri-environment schemes (AESs) to farmers in Europe and North America to make environmental improvements to their land (Donald & Evans 2006), including the reversal of previous biodiversity losses caused by agricultural intensification. The application of AESs can yield rapid recovery in farmland wildlife populations; but this only occurs where the design of management options is based on evidence of their efficacy, where they are well targeted, and where well-designed monitoring studies allow adaptive improvement of schemes over time (Aebischer et al. 2000, Vickery et al. 2004, Perkins et al. 2011). Across Europe, however, deficiencies in these respects mean that the effectiveness of AESs in driving biodiversity recovery has been mixed (Kleijn et al. 2001, 2006, Kleijn & Sutherland 2003, Berendse et al. 2004, Blomqvist et al. 2009, Wilson et al. 2009), and optimizing biodiversity conservation from agri-environment schemes remains a key policy challenge (Sutherland et al. 2006).

Because of investment in research to understand their ecological and demographic responses to agricultural change, birds offer some of the best examples of success of AESs driving population recovery at regional or national levels (Aebsicher et al. 2000, Peach et al. 2001, O'Brien et al. 2006, Perkins et al. 2011). Even among birds, however, outcomes have varied. Breeding waders (Haematopodidae, Burhinidae, Charadriidae, Scolopacidae) are of particular concern. Over the last century, the loss of around half of global wetlands, often to drainage and agricultural conversion, has been a main cause of declines of breeding wader populations (Dugan 1993, Wilson et al. 2009). In Europe, populations of several wader species breeding on agricultural land continue to decline, with Stone-curlew Burhinus oedicnemus, Northern Lapwing Vanellus vanellus (henceforth Lapwing), Ruff Philomachus pugnax, Common Snipe Gallinago gallinago (henceforth Snipe), Black-tailed Godwit Limosa limosa, Eurasian Curlew Numenius arquata (henceforth Curlew) and Common Redshank Tringa totanus (henceforth Redshank) all listed as species of unfavourable conservation status (Birdlife International 2004). Yet, studies of waders across Europe have found that current AES management often do no better than slow the rate of population declines (Ausden & Hirons 2002, Kleijn & van Zuijlen 2004, Kleijn et al. 2004, Ottvall & Smith 2006).

In the UK, long-term declines of breeding wader populations on intensively managed lowland agricultural land are severe and continuing. Intensification of agricultural grassland management through drainage, re-seeding and changes in the intensity and timing of grazing have all played their part in driving the declines by limiting nesting opportunities and reducing invertebrate food availability (Wilson et al. 2007). On arable land, Lapwings and Stone-curlews have suffered from the replacement of sparsely vegetated spring-sown crops and fallows, in which they are able to nest, by autumn-sown crops whose vegetation cover is too tall and dense for nesting (Green & Griffiths 1994, Sheldon et al. 2004, Wilson et al. 2005). In lowland England, where agricultural intensification has been most rapid and severe, formerly widespread species including Lapwings, Snipe, Curlews and Redshanks are rapidly becoming restricted to areas managed as nature reserves or land managed under the highest ‘tiers’ of agri-environment schemes (Ausden & Hirons 2002, Wilson et al. 2004, 2007). Here, management options currently aim to restrict or eliminate agricultural grazing during the wader breeding season in order to reduce nest trampling risk and maintain suitable sward structure for nesting. On lowland and marginal upland farmland in northern England and Scotland, these species still persist as birds of the wider countryside (Forrester et al. 2007), although there is now evidence of declines here too. For example, a 20-year study on marginal upland farmland in southern Scotland revealed a 77% decline of Lapwings between 1980 and 2000, associated with loss of spring-cultivated fields and unimproved grassland in favour of drained, improved grassland (Taylor & Grant 2004). More widely across Scotland, annual monitoring of wader population trends by the Breeding Bird Survey (BBS) has revealed declines of 21%, 27% and 53% for Eurasian Oystercatchers Haematopus ostralegus (henceforth Oystercatcher), Lapwings and Curlews between 1995 and 2008 (Risely et al. 2010). Redshanks are too scarce for their population trends to be monitored by BBS in Scotland, but ongoing national Atlas fieldwork suggests declines of this species in some areas (www.the-soc.org.uk/se-atlas/change/sesa2_breeding_change_0546_Common_Redshank.html). Only numbers of Snipe have increased (Risely et al. 2010).

AESs have been introduced into the Scottish countryside since 1988 through the geographically targeted Environmentally Sensitive Area (ESA) scheme, although management options designed to be beneficial to breeding waders only became available from 1993/94 (Table 1). AESs became available to farmers outside ESA boundaries in 1997 through the Countryside Premium Scheme (CPS). After 2000, both ESA and CPS were replaced by a single scheme, available across Scotland – the Rural Stewardship Scheme (RSS). These schemes targeted two categories of prescriptions designed to benefit breeding waders. ‘Grassland’ options aimed to improve the productivity of open-field, ground-nesting birds, including waders, by reducing the extent to which the field is exposed to livestock grazing and agricultural operations during the breeding season. ‘Wetland’ options provided both the aforementioned restrictions, but also created or maintained wet conditions in the field. In both cases, farmers entered into five-year management agreements on a field-by-field basis, in return for scheme payment. Since 2008, very similar management options have been incorporated within a successor to all previous schemes – the Rural Priorities element of Scotland's Rural Development Programme (SRDP; www.scotland.gov.uk/Topics/farmingrural/SRDP/RuralPriorities/Options).

Table 1. Agri-environment management options for farmland waders.

Scheme	Years available	Option description	Type for analysis	No. fields
ESA	1993–2000	Water margin grazing control	Grassland	20
ESA	1993–2000	Wetland grazing control	Wetland	3
CPS	1997–2000	Flood plain management	Wetland	8
CPS	1997–2000	‘Grassland for birds’ management	Grassland	27
CPS	1997–2000	Wetland creation and management	Wetland	6
RSS	2001–2006	Flood plain management	Wetland	3
RSS	2001–2006	Grazed grassland for birds	Grassland	38
RSS	2001–2006	Mown grassland for birds	Grassland	36
RSS	2001–2006	Wet grassland for waders	Wetland	36
RSS	2001–2006	Wetland creation and management	Wetland	4
ESA, Environmentally Sensitive Area; CPS, Countryside Premium Scheme; RSS, Rural Stewardship Scheme.

The present study revisited in 2005 farmland study areas in which breeding waders were surveyed in 1992/93, immediately before agri-environment management for these species became available. We aimed to test whether numbers of five breeding wader species (Oystercatchers, Lapwings, Snipe, Redshanks and Curlews) have shown a more positive response since 1992/93, at sites that have subsequently been included within agri-environment schemes targeted at breeding wader populations, in comparison with sites that have remained outside such schemes. We carried out this test at two different units of spatial replication; the site or 1-km square (the unit of original bird survey design) and the individual field (the unit of AES implementation). We also used information on the distribution of breeding wader-focused AES management in 2005 to assess whether such management has been directed successfully at important breeding wader sites, and to assess whether the influence of AESs on breeding wader populations on Scottish farmland is likely to be sufficient to reverse national population declines.

METHODS

Study sites and sampling design

The original surveys in 1992/93 were based on visits to:

1.	Three-hundred and seven 1-km squares selected randomly from all land classified as suitable for arable or improved grassland agriculture (Soil Survey of Scotland 1982). These squares were selected from two regional strata; the Scottish mainland plus Argyll islands (240 sites) and Orkney (67 sites) (O'Brien 1996);
2.	Two-hundred and eighty sites (of varying size and shape) identified by ornithologists as key breeding wader sites (186 sites on the mainland [O'Brien & Bainbridge 2002] and 200 sites across Orkney [Campbell et al. 1989]).

The then Scottish Executive Environment and Rural Affairs Department provided information on the location and area of all fields with agri-environment management of potential benefit to breeding waders (Table 1) as at 1 April 2005. A map of the distribution of these fields was overlaid on a map of the distribution of 1-km squares that had been randomly selected for surveying breeding waders in 1992/93, and all squares that contained ‘wader’ agri-environment management by 2005 were selected for re-survey (n = 40). Each of these 40 squares was then paired with the closest of the original randomly selected squares that did not contain such management in 2005, and this random square was also selected for re-survey. These squares are subsequently referred to as random sites. The same process was undertaken for ‘key sites’ that contained an agri-environment scheme (n = 42) and paired with an adjacent key site containing no scheme. Sixty of these 82 site pairs were then finally selected such that there were equal numbers of random and key site pairs (30 of each), split into four regions, within which a single surveyor covered all pairs of sites.

Field methods

All 60 pairs of sites were re-surveyed in the spring of 2005 (Fig. 1). Survey methods were exactly as in 1992/93 (Smith 1983, O'Brien & Smith 1992, O'Brien 1996). Three visits to each site were undertaken between 15 April and 21 June, all within three hours of either dawn or dusk. All fields on each site were visited. The observer walked to within 100 m of all points in each field and looked 200–400 m ahead, scanning with binoculars to note the distribution of all waders. The following information on the numbers and behaviour of all breeding waders was collected in each field on each visit.

Figure 1. The distribution of sites surveyed for breeding waders in Scotland in 1992 and 2005. Filled squares indicate sites with agri-environment management agreements for breeding waders under one or more options listed in Table 1; open squares indicate sites not under agri-environment management.

Oystercatchers: the maximum number of pairs recorded on any one visit to the site between 15 April and 21 June.

Lapwings: the maximum number of individuals on site (excluding flocking birds showing no signs of association with site or breeding) between 15 April and 28 May, divided by two.

Snipe: the maximum number of drumming plus ‘chipping’ birds recorded on any one visit to the site.

Curlews: the maximum number of pairs recorded on any one visit to the site between 15 April and 21 June.

Redshanks: the mean number of individuals recorded on the site between 15 April and 28 May, rounded upwards to the nearest whole integer.

Each of these bespoke methods yields a total count that is considered to be the best estimate of the number of occupied territories (breeding pairs) present in the surveyed area (Smith 1983).

Data analysis

AES effects on breeding wader numbers at the site scale

General Linear Mixed Modelling (GLMM) was used to measure the association between agri-environment management and numbers of waders, including only pairs of sites where the species concerned had been present in at least one of the survey periods (Oystercatcher, n = 58; Lapwing, n = 58; Snipe, n = 43; Curlew, n = 56; Redshank, n = 40). The following model was fitted for each wader species, specifying a log link and Poisson error distribution:

where count = the number of pairs of the species present at a site, and all predictor variables are categorical fixed effects: year = 2005 versus 1992/93; type = random site versus key site; and AES = whether or not the site included at least one field in AES management for breeding waders in 2005. The critical model term is the year *AES interaction term, which measures whether the difference in wader count between 1992/93 and 2005 differs between sites that have subsequently been influenced by AES management and those that have not. We also considered models in which the AES term was replaced by a covariate, duration, which measured the number of years that a site had been in wader-friendly agri-environment management by 2005 (value = 0 for control sites), and in which the AES term was replaced by a four-level equivalent discriminating AES sites with ‘grassland’ management options only (1), ‘wetland’ management options only (2), or both (3), in addition to control sites (0). The year *type interaction term was fitted in all models in order to control for the possibility that change in wader populations might differ between random sites and ‘key’ sites independently of agri-environment management. Site (1–120), site pair (1–60) and region (SW Scotland, E Scotland, N Scotland and Orkney, and also representing observer effects) were each considered as random categorical effects in preliminary analyses. However, only site showed a significant covariance parameter estimate, so all analyses use site as the sole random effect, thus controlling for repeat visits to sites. Denominator degrees of freedom for tests of fixed effects were calculated using the Kenward–Roger method (Littell et al. 1996). No models showed overdispersion.

AES effects on breeding wader numbers at the field scale

The same modelling was carried out at the field scale. All fields within sites that contained at least one field under AES management in 2005 were included in the analysis. In this case, individual fields were classified as either (1) under ‘wetland agri-environment management; (2) ‘grassland’ agri-environment management (Table 1); or (3) not under agri-environment management. This three-level categorical variable was now the AES effect in the model so that we could distinguish the effect of the two main types of agri-environment management directed at breeding waders. Site, region, and field were all modelled as random effects in preliminary analyses, but final analyses include field as the sole random effect as this was the only one to have a significant covariance parameter estimate.

Has AES management been focused on sites with higher wader densities?

We used logistic regression to test whether 1-km squares with (1) or without (0) AES management in 2005 varied in the presence or density of breeding waders in 1992/93. For this analysis we used the 240 squares that had originally been selected randomly to represent farmed land on the Scottish mainland. Thirty-one of these sites contained AES management in 2005 that might benefit breeding waders.

Even if AES management for breeding waders is not targeted at sites with higher breeding wader densities, because scheme uptake at the site scale is driven by farmer choice, within a site we might still expect that the fields selected would be those more likely to be occupied by waders. So, in a second test we again used logistic regression to test whether, on sites with AES management in 2005, individual fields with (1) or without (0) management differed in presence or density of breeding waders in 1992/93.

RESULTS

One-hundred and nineteen sites were surveyed in 2005 (permission to survey one site was denied). These held 887 Oystercatcher, 718 Lapwing, 277 Snipe, 528 Curlew and 201 Redshank breeding pairs.

AES effects on breeding wader numbers at the site scale

The difference in wader numbers between 1992/93 and 2005 on AES sites was more positive or less negative than on sites without such management, for all species (Fig. 2). However, this effect was significant only for Lapwings (F _1,226 =10.61, P = 0.001) and Redshanks (F _1,154 =17.80, P < 0.0001). In no case did replacement of the AES factor with the covariate, duration, improve model fit. Equally, replacement of the two-level AES factor with a four-level equivalent in order to distinguish the type of agri-environment management on AES sites showed that population change did not differ significantly according to whether grassland management options, wetland options, or both, were provided on AES sites. In summary, between 1992/93 and 2005, Lapwing numbers were reduced by 24.2% on AES sites, and by 44.1% on control sites. Redshank numbers had increased by 15.3% on AES sites, but declined by 49.2% on control sites.

Figure 2. Modelled estimates (± se) of numbers of breeding pairs of each wader species on sites with (A) and without (C) agri-environment scheme (AES) management. Numbers refer to years: 1992 (92) and 2005 (05); species for which the change in numbers between 1992 and 2005 differed significantly between AES and Control sites are indicated by asterisks (**P < 0.01; ***P < 0.001).

AES effects on breeding wader numbers at the field scale

Table 2 shows the number of fields in ‘grassland’ or ‘wetland’ agri-environment management (see Table 1), or not under agri-environment scheme management, and occupied by each wader species in one or both survey years. The sample of fields under agri-environment management and occupied by Redshanks (n = 24) and Snipe (n = 13) was small. Nonetheless, Lapwings, Snipe and Redshanks all showed changes in numbers between 1992/93 and 2005 that differed significantly between the three field types (Fig. 3). Lapwings declined by 65.2% on ‘wetland’ AES fields and by 24.8% on control fields, but remained stable in numbers on ‘grassland’ AES fields (F _2,736 = 11.74, P < 0.0001). Snipe declined by 34% on ‘wetland’ AES fields, but increased by 24.8% on control fields, and over four-fold on ‘grassland’ AES fields (F _2,232 =3.66, P = 0.027). Redshanks declined by 49.8% on ‘wetland’ AES fields and by 13% on control fields, but increased by 170% on ‘grassland’ AES fields (F _2,266 = 6.35, P = 0.002).

Table 2. The number of fields in each of the main agri-environment scheme (AES) management categories (grassland and wetland), together with the number within which a breeding territory of each wader species was recorded in at least one of the survey years.

Type	OC	L	SN	CU	RK	Total fields
Grassland	36	29	7	36	9	72
Wetland	30	21	6	24	15	43
Non-AES	445	322	106	304	113	1015
OC, Oystercatcher; L, Lapwing; SN, Snipe; CU, Curlew; RK, Redshank.

Figure 3. Modelled estimates (± se) of numbers of breeding pairs of each wader species on individual fields under ‘grassland’ agri-environment scheme (AES) management (G), ‘wetland’ AES management (W), and not under AES management (C) in 1992 (92) and 2005 (05). Species for which the change in numbers between 1992 and 2005 differs significantly between ‘grassland’, ‘wetland’ and control fields are indicated by asterisks (*P < 0.05; ** P < 0.01; *** P < 0.001).

Has AES management been focused on sites with higher wader densities?

Only 12.9% (31/240) of the 1-km squares (which represented a random sample of Scottish mainland farmland selected in 1992/93) supported AES management for waders in 2005. These sites contained between 5% (Snipe) and 20% (Redshanks) of the breeding waders recorded in 2005. However, there was no significant effect of either the presence (P = 0.15–0.87 across the five species) or number (P = 0.15–0.69) of waders of any species in 1992/93 on the probability that a 1-km square would contain AES management for breeding waders in 2005.

Of the total field area of the 31 1-km squares with AES management for breeding waders in 2005, only 5.8% was managed under the AES management options listed in Table 1. These fields contained between 0% (Snipe) and 16.7% (Redshanks) of the breeding waders occupying these squares in 2005. Logistic regression analyses for each wader species found that the probability that a field has entered AES management by 2005 was weakly positively related to the occurrence of Lapwings (P = 0.024) and Curlews (P = 0.030) and the number of Curlews (P = 0.043) in 1992/93.

DISCUSSION

The difference in breeding wader abundance between 1992/93 and 2005 was more positive on sites that had come under AES management than on sites which remained outside AES, although taking into account the duration and type of AES management did not improve model fit. These differences were marked and statistically significant for Lapwings and Redshanks, and occurred despite the fact that there was no evidence that AES management had been successfully targetted at breeding waders. However, AES management was associated only with modest population increase for Redshanks, and a reduction in the magnitude of decline for Lapwings. At the scale of individual fields, there was strong evidence for Lapwings, Redshanks and Snipe that the implementation of ‘grassland’ options was associated with more positive outcomes than the implementation of ‘wetland’ options, although the cases where this translated into population increase (Redshanks and Snipe) were based on small sample sizes (Table 2). Moreover, for these three species there was evidence that implemenation of AES ‘wetland’ options was associated with worse outcomes than leaving those fields out of AES management. At the field level, there was weak evidence that fields that were under wader-friendly AES management by 2005 had been more likely to hold breeding Lapwings and Curlews in the 1992/93 survey than fields which remained outside AES.

These results show that agri-environment management designed to benefit breeding waders is capable of slowing, and in some cases reversing, losses of breeding waders on the fields where it is applied. However, at the field scale these benefits appear limited to those ‘grassland’ options which simply limit grazing and agricultural operations on grass fields during the wader breeding season rather than also seeking to raise water levels. This may seem counter-intuitive given the very strong evidence from other studies of positive impacts of restoring high water tables for breeding grassland waders (Verhulst et al. 2007, Eglington et al. 2008). One possible explanation may be that fields entered into ‘grassland’ options tend to be on the more productive areas of farms. Such fields are returned to active grazing and other agricultural management each year, as soon as the seasonal restrictions end. By contrast, many fields entered into ‘wetland’ options may be on less productive, marginal areas of the farm and suffer from gradual, cumulative reduction in agricultural activity over the course of the 5-year management agreement, even outside the period of seasonal restriction. Vegetation development on such fields would be likely to render them progressively less attractive to breeding waders. The problem may be exacerbated if farmers fence off the wettest areas (fencing costs are eligible payments under these options) so that livestock are unable to graze them. Although the current study is unable to test this explanation in more detail, ongoing inspection and verification arrangements for the relevant grassland and wetland management options in SRDP need to be sufficient to guard against gradual succession to rank vegetation cover over the course of agreements, at least at sites where breeding waders are a target of management. At the farm scale, population changes on AES sites did not differ according to whether grassland, wetland or both categories of management option were deployed. This suggests that even though fields under wetland option managment did not themsleves support more positive population changes than control fields, they may have been beneficial to waders elsewhere on AES sites, perhaps through providing suitable cover and foraging habitat.

More generally, these results illustrate the importance of ongoing monitoring of the efficacy of AES management options as the basis for their adaptive improvement over time (Whittingham 2007; Memmott et al. 2010; Perkins et al. 2011). However, whether even the most skilful and well-targeted implementation of these options will be suffcient to reverse the national population declines of breeding waders depends on the scale over which implementation takes place (Wilson et al. 2010). If we take the example of Lapwings, then the decline of this species recorded by the Breeding Bird Survey (BBS) (27% between 1995 and 2008; Risely et al. 2010) equates to a rate of 2.4% per annum. Given that, for an assumed closed population:

where λ_w = the per annum growth rate of the population outside AES management; λ_a = the per annum growth rate of the population benefitting from AES management; and p = the proportion of the population that must benefit from AES management in order to achieve overall population stability (i.e. λ = 1), Fig. 4 shows how the value of p varies with that of λ_a, assuming that λ_w = 0.976. Our data show that in 2005, only 12.9% of the random sample of farmland 1-km squares supported any AES management designed to benefit breeding waders (supporting 13.3% of the recorded Lapwing territories), and only 5.8% of the field area within those squares was managed in this way (supporting 1.8% of the Lapwing territories). These levels of provision (<1% of the farmland area, and <2% of Lapwing territories) are far below what would be required to reverse the decline of the population of Lapwings in mainland Scotland assuming any realistic rate of population increase (Fig. 4). An annual population growth rate of 2–3% under appropriate AES management would require targetting 44–54% of the Lapwing population in order to offset declines elsewhere and stabilize the overall population trend.

Figure 4. Relationship between annual population growth under agri-environment scheme management and the proportion of the overall population that must benefit from this management in order to achieve overall population stability. Example plotted is for Lapwings with an assumed constant rate of decline of 2.4% per annum, amounting to 27% over 13 years.

Current agri-environment provision for breeding waders within SRDP certainly provides the potential to deliver more effectively targeted and comprehesive packages of management directed at breeding wader populations. As well as the ‘grassland’ and ‘wetland’ management options that were the focus of this study, applicants are now encouraged to package these together with other measures including late cutting of meadows, natural flooding of floodplain grasslands, and predator management (www.scotland.gov.uk/Topics/farmingrural/SRDP/RuralPriorities/Packages/FarmlandWaders; accessed 1 September 2010); the latter now known to be effective in increasing productivity and numbers of some breeding waders (Lapwings, Curlews and European Golden Plovers Pluvialis apricaria) in grazed, upland systems (Fletcher et al. 2010). Moreover, initiatives such as the Bird Conservation Targetting Project (www.bto.org/birdtrack/bird_recording/bctp/index.htm; accessed 1 September 2010) provide regularly updated information on distributions of farmland-nesting waders and other species of high conservation priority to allow both farmers and scheme administrators to ensure that individual agreements are targetted where they are most likely to be effective. At present, there is no programme to monitor biodiversity responses to agri-enironment intervention under SRDP and data are needed urgently to assess whether this potential is being realized, and whether agri-environment measures are now being implemented at a sufficient scale, in the most appropriate combinations and with sufficient spatial targetting on key remaining wader populations, to be able to halt declines.

ACKNOWLEDGEMENTS

This project would not have been possible without the support of the Scottish Executive Environment and Rural Affairs Department, who provided information on the type, location and duration of the various agri-environment schemes in place in Scotland between the two survey periods. The experimental design would not have been feasible without this information. Many thanks to Rosemary Setchfield, John Dyda and Dave White for undertaking the fieldwork, and to Rosemary Setchfield for inputting the data. Most importantly, many thanks to the many farmers and landowners who allowed us to access their land in order to collect the information, and with whom we had many discussions over possible reasons for declining wader populations.