Changes in the status of waders breeding on wet lowland grasslands in England and Wales between 1982 and 1989

Transcription

1 Bird Study ISSN: (Print) (Online) Journal homepage: Changes in the status of waders breeding on wet lowland grasslands in England and Wales between 1982 and 1989 M. O'Brien & K. W. Smith To cite this article: M. O'Brien & K. W. Smith (1992) Changes in the status of waders breeding on wet lowland grasslands in England and Wales between 1982 and 1989, Bird Study, 39:3, , DOI: / To link to this article: Published online: 24 Jun Submit your article to this journal Article views: 333 View related articles Citing articles: 17 View citing articles Full Terms & Conditions of access and use can be found at

2 Bird Study (1992) 39, Changes in the status of waders breeding on wet lowland grasslands in England and Wales between and 1989 M. OBRIEN* and K.W. SMITHt *Britis h Trust for Ornithology, The Nunnery, Thet ford, Norfolk 1P24 2PLI, UK froyal Society for the Protection of Birds, The Lodge, Sandy, Beds SGI9 2DL, UK A randomly selected sample of sites surveyed in 1982 for the Breeding Waders of Wet Meadows survey was re-visited in 2989 and information on numbers of breeding waders collected. This showed that over this period numbers of Lapwings Vanellus vanellus had declined significantly by 38%, whilst Oystercatchers Haematopus ostralegus had increased by 56%. There was no evidence of a change in numbers of Snipe Gallinago gallinago, Redshank Tringa totanus or Curlew Numenius arquata. In both 1982 and 1989 a high proportion of all Snipe and inland breeding Redshank occurred on nature reserves. Numbers of Snipe and Redshank (but not Lapwing) were less likely to have declined on reserves than on unprotected sites. A report to the British Trust for Ornithology L and drainage and conversion to arable cultivation have led to the loss of large areas of damp lowland grasslands in England and Wales over the last few decades. During the 1970s about 8000 ha of damp grasslands were drained annually.' In 1982, the British Trust for Ornithology (BTO), in co-operation with the Royal Society for the Protection of Birds (RSPB) and the Nature Conservancy Council (NCC) organized a national survey of Breeding Waders of Wet Meadows throughout England and Wales. 2 The objectives were to identify all areas of damp lowland grassland in England and Wales and to count the numbers of breeding waders using them. The species included were Lapwing Vanellus vanellus, Snipe Gallinago gallinago, Redshank Tringu totanus, Curlew Numenius arquata and Oystercatcher Hoematopus ostralegus. The results showed that the numbers of both waders and sites were *Present address: RSPB, The Lodge, Sandy, Beds SG19 2DL, UK. lower than expected, and that a high proportion of breeding waders were concentrated in a few exceptional areas. The only widely distributed species was the Lapwing. Subsequent monitoring of these wader populations was considered desirable, so in 1984 the Breeding Wader Monitoring Scheme (BWMS) was initiated by the BTO and the Wader Study Group (WSG) to provide information on annual fluctuations and trends in the populations of lowland grassland breeding waders. 34 The results of the scheme suggested that numbers of some wader species were still declining. 5 In addition, in 1987, an RSPB survey of the Somerset Levels found that total breeding wader numbers there had fallen by 48%, and that Snipe numbers had fallen by 58% in the 4 years since the last survey in It was against this background that in 1989 the BTO, with funding from the RSPB, conducted a survey of a randomly selected sample of all sites first covered in the 1982 survey. This report describes the changes in the lowland wet grassland breeding wader populations of England and Wales between 1982 and

3 166 M. O'Brien and K.W. Smith The effectiveness of site protection in maintaining breeding wader numbers over the seven year period is reviewed. The emphasis of this paper is on breeding populations of Lapwing, Snipe and Redshank these being the most widespread species on lowland wet grasslands in England and Wales. METHODS Selection of sites For the 1982 survey, damp lowland grass was considered to be any area of grassland subject to freshwater flooding or waterlogging such as floodplain grasslands, coastal grazing marshes, washlands as well as isolated pockets of poorly drained grassland. 2 Coverage was arbitrarily restricted to land below 600 ft (183 m) in altitude. No attempt was made to survey any areas subject to regular tidal flooding. 2 In the original survey, a total of 1398 lowland grassland sites was surveyed. Eighty-four per cent of these sites were surveyed in 1982, the remainder covered in 1981 and For the purposes of this paper, it has been assumed that there were no changes in breeding wader numbers over the period , and a starting point of 1982 has been assumed. A stratified random sample of the original 1398 sites was selected. The sites were split into 15 groups based on their location and status. Five regions were used corresponding to the old water authority (now water company) boundaries as shown in Fig. 1. These regions, based on Water Authority boundaries, were used because each covers whole catchments, so there was no risk of sites lying across regional boundaries. All sites within each of these regions were further sub-divided by their protection status. (a) Non-reserves with no breeding waders present in 1982: 'no wader sites' (20% of the 1982 sites were in this category). (b) Non-reserves with breeding waders present in 1982: 'wader sites' (the majority of sites). (c) Reserves managed by the RSPB, NCC or WWT (Wildfowl and Wetlands Trust): 'reserve sites' (33 sites). All 1398 sites from the 1982 survey were classified into one of these 15 groups. It was considered that the surveying of 300 sites Figure 1. The 5 regions of England and Wales as used in this paper. The boundaries follow National Rivers Authority boundaries as outlined in Methods. would be a feasible undertaking and would provide sufficient information on changes in wader numbers. For each region all 'reserve sites' were surveyed together with eight 'no wader sites' and approximately 25% of the 'wader sites' (Table 1). Changes in wader numbers within each group were calculated and applied to the original 1982 data to provide an estimate of the overall change. Detailed surveys, undertaken by the RSPB, of the Somerset Levels in and the Norfolk Broads in meant that data on changes in wader numbers were already available for these large areas of lowland grassland. Results from these surveys were considered to be sufficient to describe the change in numbers within these 2 areas so sites within these areas were not selected. To analyse the impact of statutary wildlife protection status on the changes in wader numbers, the sample data were regrouped into 3 protection categories: unprotected; Site of Special Scientific Interest (SSSI); and reserve. If 50% or more of a site was within an SSSI then it was classed as an SSSI. All sites that were both SSSIs and reserves were categorized as re-

4 Lowland grassland breeding waders 167 serves. Data on the extent of SSSIs were provided by the NCC, with further checks made against records held by RSPB regional offices. Details of data collection The survey was carried out mainly by amateur fieldworkers organized through the national network of BTO Regional Representatives. Observers were asked to survey all the fields within one of the randomly selected sites. Data were collected on a field by field basis, in contrast to 1982 when records were only available by site. In all other respects the surveys were directly comparable. Three copies of 1:12500 map of the site were provided, together with sufficient visit cards to record information on each field. Observers were asked to walk to within 100 m of every part of each field within the site and to complete their survey between dawn and hours. Three visits were made, one in each of the periods 15 April-30 April, 1 May-21 May and 22 May-18 June. Successive visits were separated by at least a week. Cold, wet or windy weather conditions were avoided. The location, movement and behaviour of all wader species were recorded onto the field maps using standard codes. Each bird was allocated to a single field the first field in which it was recorded, or, if the bird was first ob- served in display flight, the field at the centre of its display flight. At the end of the visit, summary information was transferred to the visit cards using the following criteria for each species: Lapwing and Redshank the total numbers of birds seen in each field; Snipe the number of drumming plus chipping birds in each field; other species the number of pairs (where 'pairs' = (paired individuals/2), displaying birds, nests or broods and other single birds not in flocks) assigned to each field. Analysis of wader numbers Since the results of the 1982 survey were published there have been a number of studies of the relationship between census results and the numbers of nesting pairs which allow the interpretation of results to be improved. For Lapwing is has been found' that both members of a pair were present at the nest site during the late stages of incubation. Hence halving the maximum number of individuals recorded on site between mid-april and late May gives the best estimate of total number of pairs. For Redshank the mean density of individual birds (excluding flocks) counted between mid-april and late May was correlated with peak nest density Fortunately, in the Table 1. Number of sites surveyed in 1982 and selected in 1989 Sites with no waders Sites with waders Region SOS RSS SAS SOS RSS SAS RES SSSI E. Anglia Midland North South SW & Wales Total An additional 13 sites in the Somerset Levels and 66 sites in the Norfolk Broads were excluded from the selection as these areas had been surveyed previously. SOS: Number of sites originally surveyed in RSS: Number of sites randomly selected from 1982 database. SAS: Number of sites actually surveyed in RES: Number of sites owned by RSPB, NCC or Wildfowl and Wetlands Trust (Reserves). All sites surveyed in both 1982 and SSSI: Site of Special Scientific Interest. Includes some randomly selected sites but not reserves. All sites surveyed in both 1982 and 1989.

5 168 M. O'Brien and K.W. Smith 1982 survey, the total numbers of birds present on each visit were recorded, so a revised estimate of the number of breeding pairs of Lapwing and Redshank could be calculated for the comparisons in this paper. The revised procedures increased estimated numbers of Lapwing by 34% from 6869 to 9186 breeding pairs, and Redshank by 35% from 2482 to 3361 breeding pairs for all sites in the 1982 survey. For other species the methods used to calculate the number of breeding birds were the same as in For Snipe the maximum number of drumming birds between mid-april and late- May was used, whilst for all other species, the maximum number of pairs on any one visit was taken. Methods of analysis Changes in wader numbers between 1982 and 1989 were calculated using the procedure employed in the Common Birds Census. 14 Confidence limits were calculated using a bootstrap method. 15 For each species the changes in numbers of randomly selected nonreserve sites were calculated for each region. Changes in the overall wader populations of lowland wet grasslands between 1982 and 1989 were calculated by multiplying the number of waders recorded in 1982 in each 'region*status' group by the percentage change recorded on those sites that were surveyed in 1989 in that group. The numbers recorded for the Somerset Levels' and for the Norfolk Broads 10 were added to the totals for SW & Wales and the East Anglia regions, respectively. An assessment of the effectiveness of protection afforded by reserve and SSSI status was made by calculating the change in wader numbers on each site and comparing this across the protection categories. The change in numbers on each site was calculated using the Index SITE INDEX = W89/(W82+W89) where W82 and W89 = number of waders present in 1982 and 1989, respectively. This index gives values of 0.5 when there was no change in wader numbers between the two years, 0.0 when waders were present in 1982 but not 1989, and 1.0 when waders were absent in 1982 but present in Each site was then classed as declining (Index = 0 to < 0.35), stable (Index = 0.35 to 0.65) or increasing (Index = > 0.65 to 1) and a 3*3 contingency table drawn up comparing change with protection status. A likelihood ratio x2 (or G) test was then used to test the possibility that there were differences between protection categories in trends in wader populations. If a significant result (P < 0.05) was obtained then a simultaneous test procedure (STP) using a G-test was used to compare all pairs of protection categories. To be significant at the 5% level these G values must exceed 9.49 (x2 with (3-1)*(3-1) degrees of freedom) in the STP tests. 15 RESULTS How representative were the sites surveyed? Although the sites to be surveyed were selected at random, Table 1 shows that data were received for only 73% of those selected. Were the sites surveyed in 1989 representative of lowland grassland sites in general? It is possible, for instance, that site data were only received from sites with waders present, thereby biasing the selection of sites actually surveyed. To test for such a difference, Mann Whitney U-tests 16 were used to compare 1982 wader numbers on sites selected and surveyed in 1989 with those selected but not surveyed. Of 25 tests (5 wader spp. in 5 regions), 24 showed no significant difference at the 5% level. Redshank numbers were higher on surveyed sites within the Midland Region, but in all other cases there were no differences in wader numbers between surveyed and unsurveyed sites. Given 25 tests one significant result would be expected by chance. Hence there is no reason to believe that incomplete coverage will have biased the results of this sample survey. Changes in numbers of Lapwings, Snipe and Redshank Lapwings were the most numerous and widespread breeding wader on lowland wet grasslands in A significant decline between 1982 and 1989 was recorded in 3 of the 5 regions (South, SW & Wales and Midland) (Table 2a). A significant decline was also recorded on reserves (Table 2a) in the East Anglian region, whilst numbers on reserves in the Northern region increased significantly.

6 Lowland grassland breeding waders 169 } m N U 1 + ó O -9., 0') I H O I m oz, M,r c-+,)% 1 t + 1 I I +.ND r-,, +1 n N tr) + + I + cs O O GO rm Ir)O 00 N O NN IS) r In I-. r., -+ M cn ti r. C N N r Cr O -n O N Cr),--I O M N,0 rh r-+ O O N r n -,-' N N. In r. r. Gy co R d..t c3-n N m,-i Ifj CO O N N O O O O O O p O O O O N L. rh I-. Nti Cr) r-n M N +V',D M W t\ <\ <\.0 I I n,..,....., _..._.. nn. nn o O Cr N N I.(} N O, N N N N d+ N h Ir) 1 N 1 N p I I I I I + N rn hr,,t.,z)in N th [` In,-N- N NCr) O LC) N N CV +, 2 mt*! Cr M Cr)In N N N m N N M,--I N N N a0 N y v o co y m N,C) L N 00 N N d' R :49 a al II) m In CS, C fn Cr) 00 N Cr) m,-+ r.,-. V' 00 N N N N r-' O fl A R ` Á O v c. Cr) <M In N O r r-a O O v o R R v o 00 Q.4.0,ó ó,,v...5 `',ei' W Zcfi cn W zcnc!) x A b 4,,,,$ C4 W tzcnc/3)

7 170 M. O'Brien and K.W. Smith Vn C, to COon r, 01o tn m o rn C. in r, tn e- 1,O N N a\ CO O`O N to O cc t\ r, r, tt) N GO O r, r, n CO O r,n -+ e-, tf) r+ m to O r, COO a, a1 LOCO0 On CO OD o r, N U, CO CO CD r, r, N cc to N n h at, C San oo r, m -1 " ti m N NCO0 r', am, r, CV) COto. N CO N Cr CO n i\ N O N N Ns CO COto t[) O tri d+ N us N a fo tm to N N d,. LC)S, r, CO CO CO CO M O Crto m to r, D 71., N o M O COD O r, N in satt'i '-( 4< to to N to M n t Cr tp) u') to N r,.-, o r, W -, o I-4 A C tf) N 4o.-+ N M N tf) r, O r, (n N CA CO N VD4D [, N to to AD N to tn Cr Nr, to 4D CV tb CD o C CS r, o0 t+) r,t+') r,,-, N. CS VD CC,CTCt+ O.-, o r~ to o r,,-. r, r, t+) r N o t*a CT to O C O 10 to N CO N N N,r, rn,m co CO O rni W r< r,.-n, rn, N r- N C7 0 CS o C.N, o N, 11, C. Cr,M CO C.o a. tn Cr rn N GO to N N Cr 4o N a, a1 O CT a, GO,O to 0 CS N 00 N 4D N 00 N r, ao CT r, p to O.-, tn O tn tn C, N O to to to a to a O N 00 N co CS GO N O co N N O N l\,o O tn o to N VDtf} CM 0 N mn CM r-i tn fy).-+ Cl+ cr, O C n N C N a, co as o S, Ns CO CDN C, to to d, I,D N co IS.-, N M.1, 10 a\ 0 0N Nin to ap, S. N r~ 1n tt) 'CO N If)to r, S.47 r, n 11. to r,n r, r-, to I + r-, N NN to Ct, CS,N N a,a, N r, tn m CT 0 m N, tn O,to r,,o ta 00,o tn 46 Cl tt tn,ct r, O, 4D 4D N a, O, r, r,,c) O r, N N,O M, N 6! N v > u) v u v tn á N NJ y N 6N N v á Y b 04N :4 al 4 t S s 5,, : ' -a-4 H :Z H ó Z H ó2 H z t n H ó -8 x Z n u) U r, ó a C W CO v dod o N o N U.á a Z G N Q 'O C +m 42' G E an0 T m v _ g > -C H o L' o. x E CZ g n m O 00 y v o v a II T >< ".. W X v v zq Q L a O [y. OD "c1; y v n m cn W 77 C O m Cd > -Ort,., p v O O p ac C C?2, a C v, Á m E x m `J + Fo ó c n II o cc O 77 c]n 2 c "a m bn E oo a, aj V Oa N 3 H N J v m Ex Á x 3 C W o,s) N.,

8 Lowland grassland breeding waders 171 A total of 14 pairs of Lapwing were recorded on sites which held no waders in Numbers on the Somerset Levels declined by 51% from 256 pairs in 1983 to 129 pairs in 1987 whilst numbers on the Norfolk Broads increased from 227 pairs in 1982 to 246 pairs in 1988, an 8% increase. In 1982 Snipe were recorded on 30% of all sites with a total of less than 2000 drumming birds.'' Only one region showed a significant decline in numbers between 1982 and 1989 (Southern region) (Table 2b), although total numbers dropped in the remaining regions. No significant changes occurred on reserves. 147 drumming Snipe were recorded on the Somerset Levels in 1983 which, by 1987, had declined to 62 (-58%). Numbers surveyed in the Norfolk Broads increased from 39 drummers in 1982 to 64 in 1988, a 64% increase. Re-analysis of the 1982 data showed that 3361 pairs of Redshank were recorded on 34% of the sites (see Methods). 2 Only one region showed a significant change in numbers (Table 2c) a decline of 51% in the Midland region (but note non-representative nature of sample of sites from Midland region compared with 1982 database). 'There was no significant change in numbers on reserves although the increase in the North was very close to significance. Sixty-one pairs of Redshank were recorded on the Somerset Levels in 1983 and 41 in 1987 a decrease of 33%. In the Norfolk Broads numbers increased from 47 pairs in 1982 to 57 in 1988 a 21% increase. Estimated total numbers of waders breeding on wet lowland grasslands in 1982 and 1989 are shown in Table 3. Overall numbers of Lapwings decreased significantly by 38% between 1982 and Most Lapwings were found in the East Anglian and Northern regions in There was no significant overall change in Snipe numbers on lowland grassland sites, although, when information from the Somerset Levels is included, numbers declined significantly in the SW & Wales region. Of all Snipe, 60% occurred in the Anglian region and, overall, 40% were on reserves predominantly the Ouse Washes, Nene Washes and the Derwent Ings. Redshank have shown no change in breeding numbers on lowland grasslands since 1982, The easterly distribution noted by Smith 2 was still apparent and has also been noted in the distribution of Redshank on saltmarshes. 13 Almost 70% of all lowland grassland Redshank were in the Southern and Anglian regions. Site protection and changes in wader numbers Table 4 shows change in numbers of waders in relation to different levels of site protection. For Lapwing (Table 4a) different levels of production were not associated with any differences in trends in wader numbers. For Snipe and Redshank however, there was evidence of a relationship between site protection and changes in breeding wader populations. There were large declines in Snipe on more than 40% of unprotected sites but only on 15% of reserves. Stable numbers of Snipe were maintained on over 50% of reserves but only on 20% of unprotected sites. For Redshank over 50% of the unprotected sites showed a decline in numbers compared with less than 25% in reserves, whilst over 50% of reserve sites remain stable compared with just 20% of unprotected sites. Changes in numbers of curlew and oystercatcher In the 1982 survey a total of 750 pairs of Curlew were recorded on a total of 15% of the sites. The majority of these birds were in northern England and Wales with 2 or 3 isolated pockets in southern and central England. Although the number of sites with breeding Curlew was low, the data suggests that numbers have increased in the North and declined elsewhere (Table 5). Small numbers of Curlew breed on the Somerset Levels, declining from 31 to 25 pairs9 between 1983 and The Oystercatcher is increasing as an inland breeding bird. In the 1982 survey pairs were found on 11% of the sites. Although numbers are still small, there was an increase in 3 of the 5 regions (East Anglia, Midland and Southern) between 1982 and 1989 (Table 6). Estimated total numbers suggest an overall increase of 56% which is particularly marked in the East Anglia and Southern regions. A 65% increase in numbers on the Norfolk Broads between 1979 and was followed up by a further 52% increase between 1982 and

9 172 M. O'Brien and K.W. Smith Table 4. Effect of different levels of protection on index of change in wader numbers Number of sites with Decline Stable Increase Total (a) Lapwing Reserves SSSI Unprotected Total G-test = 2.54, P > 0.05, DF = 4 (b) Snipe Reserves SSSI Unprotected Total G-test = 13.97, P < 0.01, DF = 4 STP Unprotected vs. Reserve G-test = 13.0 (P < 0.05) Unprotected vs. SSSI G-test = 0.6 (P > 0.05) SSSI vs. Reserve G-test = 6.7 (P > 0.05) (c) Redshank Reserves SSSI Unprotected Total G-test = 14.83, P < 0.01, DF = 4 STP Unprotected vs. Reserve Unprotected vs. SSSI SSSI vs. Reserve G-test = 12.5 (P < 0.05) G-test = 2.5 (P > 0.05) G-test = 4.0 (P > 0.05) The G-test measures the effect different levels of protection have on the proportion of sites in each of the categories of change. If protection status has little or no effect then the G-statistic will be low. See Methods for more details. Table 5. Changes in numbers of Curlew between 1982 and 1989 Non-reserves Reserves Estimated numbers Region East Anglia Midland North South SW & Wales Total Change (%) 5% +83% 1% Numbers in italic indicate a significant change (P < 0.05) compared with 1982.

10 Lowland grassland breeding waders 173 DISCUSSION The number of Lapwing breeding on wet lowland grasslands declined significantly between 1982 and 1989 a fall of 38%. In addition, the number of sites holding breeding Lapwings has declined. This decline has been greatest in the south and least in the north. Nevertheless, Lapwing remains both the most common and frequently occurring breeding wader on lowland grassland. Table 7 shows the change in numbers of Lapwing as recorded by the BTO's annual monitoring schemes. The variation in the percentage change may reflect the variation in the location of the sites contributing to each of the schemes. The Common Birds Census (CBC) indicates 14 that Lapwings have declined in the south, at least since the mid-1960s, but that this was partially counterbalanced by increases in the north, and particularly the north west, up until the mid-1970s. Subsequently the index dropped sharply. 14,19 Snipe have shown no significant change in numbers between 1982 and The confidence interval around a decline of 12% (-35% to +36%) indicates a high degree of heterogeneity in the changes on each of the sites. This is apparent from Table 4b numbers have declined on 46 sites, increased on 44 sites and remained stable on just 33 sites. The Breeding Wader Monitoring Scheme (BWMS) has been the only scheme to monitor changes in Snipe numbers between 1982 and 1989 (Table 7). The CBC index recorded a decline between 1968 and after which numbers were so low that an index could not be calculated. Comparison of the Snipe distribution in the Atlas of Breeding Birds with that found in shows a marked reduction in the number of occupied 10-km squares between the two surveys, par- Table 6. Changes in numbers of Oystercatcher between 1982 and 1989 Non-reserves Reserves Estimated numbers Region East Anglia Midland North South SW & Wales Total Change (%) +66% +5% +56% Numbers in italic indicate a significant change (P < 0.05) compared with Table 7. Change in numbers of breeding waders between 1982 and A comparison of the change estimated by the present survey with the annual monitoring schemes coordinated by the BTO Species Present survey BWMS CBC WBS Lapwing 38% (-52 to 18) 11% 50% 4% Snipe 12% ( -35 to +36) 0% 65% Redshank 0 % ( -36 to +50) +15% 30% 1% Curlew 1% 16% +1% Oystercatcher +56% 22% +46% BWMS: The Breeding Wader Monitoring Scheme, (Smith & O'Brien, unpublished data). CBC: Common Bird Census (J. Marchant, pers. comm.). WBS: Waterways Bird Survey, (S. Carter, pers. comm.).

11 174 M. O'Brien and K.W. Smith titularly in southern England. Mitchell' ' showed that, in Sussex, although the number of Snipe had not fallen substantially between and the number of sites holding them had declined. A reduction in range between and has also been reported in North East Scotland 22 suggesting that the decline in Snipe numbers is not just a southern phenomenon. Of the total numbers of Snipe estimated to be breeding on lowland grassland, 40% were on reserves, with relatively low numbers on sites with no form of protection. Numbers declined more on unprotected than protected sites. Improvement of marginal land in northern England has been shown to result in the virtual disappearance of Snipe 29 whilst the decline in numbers on the Somerset Levels has been linked to changes in the operation of the drainage pumps in the area 24 with an associated effect on soil moisture conditions which are known to be important for breeding Snipe. 12 For Redshank the survey estimated no change in numbers between 1982 and 1989 although once again the confidence intervals were very wide (- 36 to +50%). Only a quarter of the sites showed stable populations (Table 4c). Each of the BTO's annual monitoring schemes have also shown markedly fluctuating Redshank numbers, although a downward trend is suspected for both the CBC and WBS schemes. 14 The survey indicated that numbers of Curlew appeared to have declined in the south, but increased in the northern region. The numbers involved in the south were, however, very low and this apparent decline may have represented just a local change in the breeding area. Little is known about the detailed breeding habitat requirements of this species although Curlew were found to be limited to rough grazings and damp pastures in Scotland, 25 whilst in the marginal grasslands of Northern England densities were considerably higher on unimproved compared with improved grasslands. 23 The CBC suggests a constant population whilst the WBS indicates a significant increase. 14 It is likely that, because of small and probably unrepresentative samples, neither index can be relied upon. In recent years Oystercatchers have increased in numbers on lowland grassland sites, particularly in the south and east of the country. In 1982 they were associated with river banks, patches of arable land and occasionally gravel pit complexes. 2 Oystercatchers have shown a preference for improved and arable areas in Scotland í4 and are associated with improved grassland on marginal grasslands in northern England. 23 In Lancashire, Briggs26 outlined the expansion of Oystercatchers from their coastal nesting sites, onto the gravels of slow moving rivers or reservoirs and finally onto agricultural valleys some distance from any body of water. Oystercatchers have been monitored by the WBS since 1974, since when a steady increase has been maintained. Although this index is biased towards sites in northern England it is in agreement with the findings of this survey. Both Redshank and Snipe declined less on reserves than on unprotected sites. Much management on lowland wet grassland reserves aims to increase site wetness and to reduce grazing pressure or farming activities in spring for the benefit of breeding waders. 12 The length of breeding season of Snipe, and hence its breeding success, has been shown to be related to the wetness of the soi1, 12' 27 whilst Redshank are associated with areas of low agricultural intensity. 23,25,26 The presence of reserves on lowland wet grassland sites does not seem to have prevented a decline in Lapwing numbers. Lapwings select areas of bare ground or very short vegetation as nesting sites. 12 ' 29 Shrubb & Lack 18 suggest that optimal breeding conditions for Lapwing are found in areas of mixed farmland grass fields interspersed with spring tillage these conditions are unlikely to be found on reserves. Alternatively, reserve management, as outlined above, may change the vegetation community, 30 which may adversely affect the suitability of the reserve as a breeding site for Lapwing. One of the basic underlying assumptions of the sampling strategy adopted in this survey was that the 1982 survey was complete and covered all lowland wet grassland sites of significance. Although this is largely true in the south, where virtually no additional sites have come to light since the original survey, in the north coverage was less good. This would be of no consequence to the conclusions on overall trends if they were the same throughout the

12 Lowland grassland breeding waders 175 geographic range but, as discussed in the results, there is clear evidence of different trends in the north and south. The conclusions on the overall changes in wader numbers must therefore be treated with some caution. Lowland wet grasslands are only one of the habitats used by the breeding waders considered in this paper and for most species only hold around 10-20% of the total British breeding populations. 31 The trends reported in this paper therefore only apply to a minority of the population. To monitor the populations as a whole will require a strategy which recognizes the range of habitats used by the birds and samples them appropriately. This is one of the key problems underlying the disparate results of some of the CBC and WBS indices for breeding waders discussed above. Given the changes in breeding wader numbers reported in this paper there is clearly a need to ensure that the populations continue to be monitored. How should this be done in the future? The number of sites contributing to the annual monitoring scheme has declined to the point where it is probably too small and biased to be provide representative indices. The best way forward would appear to be regular, probably every 5 years, sample surveys covering the range of habitats used by these breeding waders. Baseline surveys will be required in habitats and areas not already covered to allow the design of an effective and meaningful sampling strategy. In addition to lowland wet grasslands in England and Wales such baselines already exist for Lapwings in England and Wales, 19 Redshank on saltmarshes, 13 a sample of Scottish farmland 14 and all breeding waders in Northern Ireland. 32 It may soon be possible to pool these data, fill any gaps and, for the first time, design a realistic sampling survey of our breeding wader populations. ACKNOWLEDGMENTS A survey of this kind is only possible with the help of a large number of people. We would particularly like to thank the BTO's Regional Representatives and members for organizing and collecting the information in their spare time. M. Ausden collected information on many sites with no waders. M. O'Brien would like to thank Liz Murray for much help with administration. The Breeding Waders of Wet Meadows Advisory Committee (F. Gribble, R. Fuller, R. Green, D. Stroud, C. Galbraith, R. Buisson) gave much advice at the planning stage of the survey. D. Stroud of the Nature Conservancy Council and RSPB regional officers provided information on SSSIs. J. Marchant and S. Carter provided CBC and WBS data on changes in wader numbers. Drs R. Green, R. Fuller, D. Parkin, J. Goss-Custard and an anonymous referee made valuable comments on earlier drafts. The project was funded by the RSPB. REFERENCES 1. Carter, E. (1982) Land Drainage. Farming and Wildlife Group Newsletter. Spring/Summer Smith, K.W. (1983) The status and distribution of waders breeding on wet lowland grassland in England and Wales. Bird Study, 30, Smith, K.W. (1984) Lowland wader surveys. BTO News, 131, Smith, K.W. (1984) Breeding Waders of Lowland Grasslands a new WSG/BTO project. Wader Study Group Bulletin, 40, Smith, K.W. (1985) Breeding wader monitoring. BTO News, 136, Smith, K.W. (1986) Breeding wader monitoring. BTO News, 143, Smith, K.W. (1987) Breeding wader monitoring. BTO News, 149, Smith, K.W. & O'Brien, M.G. (1989) Breeding wader monitoring. BTO News, 161, Robins, M. (1987) Somerset Moors breeding birds Unpublished Report. RSPB, Sandy. 10. O'Brien, M. & Buckingham D. (1989) A survey of breeding waders on grassland within the Broads environmentally sensitive area in Unpublished Report. RSPB, Sandy. 11. Barrett, J. & Barrett, C. (1984) Aspects of eensussing breeding Lapwings. Wader Study Group Bulletin, 42, Green, R.E. (1986) The management of lowland wet grassland for breeding waders. Unpublished report. RSPB, Sandy. 13. Cadbury, C.J., Green, R.E. & Allport, G. (1987) Redshank and other breeding waders of British saltmarshes. RSPB Conserv. Rev. 1, Marchant, J.H., Hudson, R., Carter, S.P. & Whittington, P. (1990) Population trends in British breeding birds. British Trust for Ornithology, Tring. 15. Sokal, R.R. & Rohlf, F.J. (1969) Biometry. Freeman, London. 16. Fowler, J. & Cohen, L. (1986) Statistics for

13 176 M. O'Brien and K.W. Smith Ornithologists. British Trust for Ornithology, Tring. 17. Murfitt, R.C. & Weaver, D.J. (1983) Survey of the breeding waders of wet meadows in Norfolk. Norfolk Bird and Mammal Report, 26, Shrubb, M. & Lack, P.C. (1991) The numbers and distribution of Lapwing V. vanellus nesting in England and Wales in Bird Study, 38, Batten, L.A. & Williamson, K. (1974) The Common Birds Census in the British Isles. Acta Orn. 14, Sharrock, J.T.R. (1976) The Atlas of Breeding Birds in Britain and Ireland. British Trust for Ornithology, Tring. 21. Mitchell, O. (1983) The breeding status and distribution of Snipe, Redshank and Yellow Wagtail in Sussex. Sussex Bird Report, 34, Buckland, S.T., Bell, M.V. & Picozzi, N. (1990) The Birds of North-East Scotland. North-East Scotland Bird Club, Aberdeen. 23. Baines, D. (1988) The effects of improvement of upland, marginal grasslands on the distribution and density of breeding wading birds (Charadriiformes) in northern England. Biol. Conserv. 45, , Robins, M. & Green, R.E. (1988) Changes in the management of water levels on the Somerset Moors. Unpublished Report. RSPB, Sandy. 25. Galbraith, H., Furness, R.W. & Fuller, R.J. (1984) Habitats and distribution of waders breeding on Scottish agricultural land. Scottish Birds, 13, Briggs, K.B. (1984) The breeding ecology of coastal and inland Oystercatchers in north Lancashire. Bird Study, 31, Green, R.E. (1988) Effects of environmental factors on the timing and success of breeding on common snipe Gallinago gallinago (Ayes: Scolopacidae). J. Appl. Ecol. 25, Beintema, A.J. (1986) Man-made polders in the Netherlands: a traditional habitat for shorebirds. Colonial Waterbirds, 9, Galbraith, H. (1988) Effects of agriculture on the breeding ecology of Lapwings Vanellus vanellus. J. Appl. Ecol. 25, Burgess, N.D., Evans, C.E. & Thomas, G.J. (1990) Vegetation Change on the Ouse Washes wetland, England, , and the affects on their conservation importance. Biol. Conserv. 53, Piersma, T. (1986) Breeding Waders in Europe. Wader Study Group Bulletin, 48, Partridge, J.K. (1988) Breeding Waders in Northern Ireland. RSPB Conserv. Rev. 2, (MS received 3 October 1991; revised MS accepted 12 November 1991)