BR.J.ENT.NAT. HIST.,20;2007 21 BEES AND WASPS IN THE DIVERSIFIED CONIFEROUS WOODLAND SETTINGS OF BRITISH CENTERPARCS Steven Falk Warwickshire Museum, Market Place, Warwick CV34 4SA [email protected] Abstract The potential value ofdiversified coniferous woodland settings for bees and wasps was revealed by a long-term study of four English Centerparc villages. These collectively produced a list of 259 species (over 50% of the mainland British list) including numerous rarities. Sampling at Elveden Centerparc in Suffolk produced a list of203 confirmed species making it one ofthe richest sites for bees and wasps in Britain today. The quantity and diversity of open habitats within a coniferous setting, and their management, have a major bearing on the quality ofthe bee and wasp fauna. A checklist ofmanagement techniques to encourage a diverse bee and wasp fauna in coniferous settings is provided. Introduction Following the end ofWorld War II, large expanses of heathland, moorland and broadleaved woodland (usually ancient) were planted up with serried rows of fast growing conifers in an effort to make Britain self-sufficient in timber. This planting programme resulted in some significant losses of valuable semi-natural habitat, particularly in areas such as Dorset, East Anglia, Nottinghamshire and some other heathland and moorland districts. Sadly, the quality of some of the timber was so poor that financial loss was incurred and there has been a degree ofsoul searching overcertain schemes ever since. Formanywildlife groups, including bees and wasps, there are relatively few opportunities in coniferous plantations other than within any rides, or the temporary clearings created by felling and replanting. In recent years, the Forestry Commission (particularly its Forest Enterprise wing) has declared an intention to restore broadleaved woodland, heathland and other habitats ofhigher nature conservation and amenity value within areas ofconiferous plantation (notably within Dorset and Breckland) as a contribution to the UK BAP targets for Priority Habitats such as Lowland Heathland (UK Government, 1994; UK Biodiversity Steering Group, 1995: Anon, 1998). Conifer removal to create heathland is also a target ofseveral local biodiversity action plans, including those for Dorset, Hampshire and Suffolk. But how does one go about converting a large, relatively impoverished conifer plantationinto anecologicallydiversesitesupportinga rangeofhabitatswithagood range offlora and fauna that includes rare and unusual species? In 1994, the author was commissioned to survey aculeate Hymenoptera and Diptera at Elveden Centerparc, Suffolk. This was the start ofan 11-year study that expanded to cover four coniferised Centerparc settings in widely separated parts of England. It has provided an ideal opportunity to answer the above question for bees and wasps within sites that still remain partiallyconiferised and are also subject to high levels of recreational activity. 22 BR.J. ENT. NAT. HIST.,20:2007 Methods Study Sites There are currently four British Centerparcs, all within England: 1. Longleat Forest-located at the easternend ofthe Longleat Forest-Longleat Park complex, Wiltshire (ST8342), in plantation woodland dominated by non-native conifers such as silver firs and redwoods. The ground mostly consists of compacted acidic sands and clays, though small patches ofmore calcareous soil are present. This is the hilliest Centerparc. 2. Elveden Forest, Suffolk (TF8080) (Fig. 1)-located within Thetford Forest, adjacent to Fakenheath Warren SSSI, and close to Wangford Warren which once supported inland mobile dunes. The plantation woodland is dominated by pines and ground conditions comprise mostly loose, chalky sands, producing some dry calcareous grassland, but also some more acidic areas with bracken and limited heathland. 3. Sherwood Forest, Nottinghamshire (SK6268) as its name suggests is part ofthe larger Sherwood Forestcomplexand isdominated bypinesonconsistentlyacidic, sandy soils. There is also a small block of semi-natural broadleaved woodland along the entrance drive. 4. Whinfell Forest, Westmorland (NY5227)-formerly known as the ‘Oasis’, is located within a larger block ofconiferous plantation (though less extensive and moreisolated than theothersites)which isdominated bypinesand featuresacidic soils that range from compacted sand to peat and loam. Figure 1. A typical Centerparcresidential landscape(Elveden), with villasarranged inclearings containing a pool and informally managed habitats such as grassland and scrub that can support good assemblages ofground-nesting and aerial-nesting aculeates. BR.J.ENT.NAT.HIST.,20:2007 23 The four villages are entirely, or mainly, located within land that was dense coniferous plantation prior to development. The landscaping that has taken place since follows a generally similar pattern featuring the following components: (i) Coniferous woodland-the dominant habitat, sometimes featuring the original plantation woodland in a dense, unmodified state, but increasingly being thinned to promote a better ground flora. (ii) Broadleaved woodland-best represented at Sherwood (semi=natural in character), but with some young plantation woodland at Elveden and small patches ofsecondary woodland at Longleat and Whinfell. (iii) Villa and plaza landscapes-several hundred villas typically located within thinned coniferous woodland, and each typically arranged with its rear end facing pools or linearwater features (except atWhinfell Forest which only has a boating lake); also a central plaza with restaurants and sport facilities. The villas tend to be located within informally managed areas with minimal mowing and naturalistic landscaping, whilst the plaza areas tend to have formal landscaping lacking semi-natural habitat. (iv) Waterfeatureswithinclearings-includes amainboatinglake, whichmayhave some useful surrounding habitat (e.g. Longleat, Elveden, to a lesser extent Sherwood and scarcely at all at Whinfell), plus subsidiary lakes and linear water features running through the villa landscape (except Whinfell). The linear water features are located within larger clearings that typically feature emergent and marginal vegetation grading into grassland, heathland or tall herb, scrub and surrounding woodland. In some cases the grassland has been enhanced by artificial seed mixtures, though at Elveden the seed bank has also produced an impressive natural response involving some very rare native plants such as fingered speedwell Veronica triphyllos L. and white horehound Marrubium vulgare L. (v) Wildflower meadows and other discrete grasslands-ofvarying quality, extent and character. Particularly impressive at Elveden due to the chalky sands that underlie the site, and parts of Longleat (Fig. 2). (vi) Restored heathland-as alltheplantations occuron sandy soils in districts that historically featured heathland, heathers have often established themselves naturally followingwoodland clearance. But special measures have been taken to promote the heathland at Longleat (Nockatt Coppice Nature Reserve) and Sherwood (compartment 45). (vii) Car parks-the main car parks (excluding Longleat) feature scrub, fringes of grassland and sparsely-vegetated sand (Fig. 3). These can be very productive. (viii) Banks, balancing pools and dumping areas-small areas often managed informallyand oftenacquiringaveryflowerycharacterafterafewyears-most notably at Whinfell where such areas provide the only substantial areas of flower-rich habitat within the village. (ix) Logpiles-notably at Elveden and Whinfell, though often temporary in nature (Fig. 4). Theresultisfoursitesthatcanbebestdescribed as ‘diversified coniferoussettings’. Methodology All the survey work was undertaken on a commercial basis, so the amount undertaken in a given year was dictated by contractual agreement. It typically involved two days worth of surveying of each site in a given year. Occasionally it 24 BR.J.ENT.NAT. HIST..20:2007 Figure 2. An example ofa 'created’ wildllower meadow (Longleat). With the right choice of plants and management, these can provide forage areas for many aculeates, especially where plants such as thistles, umbellifers and bird’s-foot trefoils are present. Figure 3. Many areas ofa Centerparc village can be managed to encourage wildlife such as aculeates with the right management, including verges, car parks, golfcourses and lakesides. BR.J.ENT.NAT.HIST.,20:2007 25 Figure4. Logpiles at Elveden. These are vital for aerial nesting species and informal dumping and storage areas such as these can be very flowery. would involve three, but ifbad weatherwasencountered (and much effortwas taken to avoid this), only a single day ofpropersurveyingmight bepossible. Dates ofvisits were varied from year to year in order to capture data from April to August, and occasionally later. Elveden was surveyed from 1994-2004, Longleat and Sherwood from 1995-2004 and Whinfell Forest (a late addition to the Centerparc portfolio) from 2002-2004. Dates of visits (bracketed dates refer to visits during sub-optimal weather): • Longleat: 30.iv.95, 12.viii.95, 2.vi.96, (28.vii.96), 19.vii.97, 20.vii.97, (13.vi.98), 5.vii.98, (9.X.99), (lO.x.99), b.vi.OO, 8.vi.00, (20.vi.01), 21.vi.01, (l.vii.02), 2.vii.02, 18.vii.03, 19.vii.03, 20.vii.03, 25.vi.04, 27.vi.04. • Elveden: 23.iv.94, 23.vii.94, 20.V.95, 6.viii.95, 23.vi.96, 17.viii.96, 21.vi.97, 22.vi.97, ll.vii.98, (12.vii.98), 27.V.99, 28.V.99, 24.V.00, (25.V.00), 6.vii.01, (8.vii.01), 17.V.02, 28.vii.02, 16.iv.03, 5.vii.03, 19.vii.04, 7.viii.04. • Sherwood: (24.vi.95), 30.vii.95, (28.iv.96), 13.vii.96, 30.V.97, l.vi.97, 29.viii.98, 13.ix.98, 26.vi.99, 18.vii.99, 6.V.00, 7.V.00, 3.vii.01, 4.vii.01, S.vii.Ol, 20.V.02, 21.V.02, 13.iv.03, 29.vi.03, 18.vii.04, 21.viii.04. • Whinfell: (15.vii.02), 16.vii.02, 13.vi.03, 14.vi.03, 15.vi.03, (28.V.04), 30.V.04. Survey techniques included visual surveillance of foraging and nesting habitats and careful sweeping of different habitats with a long-handled insect net. Areas surveyed included patches offlowers, sparsely vegetated areas, dead wood in sunny locationsand sunlit foliage. Special attentionwas given to theflowers that supported the biggest foraging assemblages at particular times of year, which included (depending on the site): 26 BR.J. ENT.NAT. HIST.,20:2007 Spring: grey willow (sallow) Saiix cinerea, blackthorn Primus spinosa, wild cherry Primus avium, hawthorn Crataegus monogyna, common gorse Ulex europaeus, dandelions Taraxacum spp., daisy Beilisperennis, coltsfoot Tussilagofarfara, cow parsley Anthriscussylvestris, ground-ivy Gleclwmahederaceaand rhododendrons. Early summer: cat’s-ear Hypochaeris radicata, mouse-ear hawkweed Pilosella officmarum, birds-foot trefoils Lotus spp., oxeye daisy Leucanthemum vulgare, foxglove Digitalis purpurea, bramble Rubus fruticosus agg., roses Rosa spp. and broom Sarothamnus scoparius. Mid to late summer: thistles {Cirsium and Carduus), ragworts Senecio spp., knapweeds Centaurea spp., hogweed Heracleum sphondylium, upright hedge parsley Torilis japonica, clovers Trifolium spp., vetches Vicia spp., rosebay willowherb Chameriou angustifolium, ling Callima vulgaris and specifically at Elveden: parsnip Pastinaca sativa, mulleins Verhascum spp., viper’s bugloss Ecliium vulgare, field scabious Knautia arvensis and mignonettes Reseda spp.. Late summer-early autumn: hawkbits Leoutodon spp., scentless mayweed Tripleurospermum inodorum and late flowers from the previous category. Persistence and timingwere very important forproducinggood lists. Somemining bees, for example, peak very early in the year when the sallow and blackthorn are in blossom. Einding bees with very narrow foraging habits required careful surveillance of specific flowers, including mignonettes for Hylaeus signatus (Panzer), field scabious for Andrena liattorfiana (E.) and heathers for Andrena fiiscipes (Kirby) and Colletes succinctus (L.). Eor genera containing species that were indistinguishable in the field, such as Spheeodes, Lasioglossum, smaller Andrena, Crossocerus, Pem- phredon and Chrysis, reasonably-sized series were obtained from as many different parts of the site as possible for critical checking under a microscope. A day in the field typically constituted 5-6 hours, including a short lunch break and walks between different compartments. Three parameters were then used to assess the quality ofthe assemblages present: (i) Species richness. This figure was the total number of bee and wasp species recorded per site during the survey period. (ii) The presence of rare species (Red Data Book and Nationally Scarce species). Thesewere taken from Falk (1991). However, to allowforthefactthat someof these gradings are known to be misleading, an asterisk has been placed against the obviously misgraded species in the Appendix and a bracketed re-grade suggestion that is more realistic given afterwards. A rarity score (RS) was evaluated for each site by assigning 100 points to Red Data Book species and 50 points toNationally Scarcespecies (following Ball, 1986). Regional scarcity, which can be used in addition to the above grades, could not be assessed easily for these sites due to the lack of county data to hand. The rarity score was divided by the total number of species at a site to produce a species quality index (SQI) which is stated to even out variable recording coverage across multiple sites. These calculations were made foreach site following adjustment for obvious misgradings. (iii) The quality of certain habitat-associated assemblages. A variety of habitat- ^ linked insect assemblages can be used to compare site quality within certain defined parameters. At the sites studied here, these included: • Acidophilous species-speeies strongly associated with heathland and acid grassland, not simply sandy soils, the presence of which can be used for monitoring the quality of any existing heathland or of any heathland - BR.J.ENT.NAT. HIST.,20: 2007 27 restoration (species highlighted as ‘Acidoph’ in the Appendix). It includes two species specifically requiring heathers for foraging, Andrenafuscipesand CoUetes succinctus. • Calcicolous species-species strongly associated with chalk or limestone habitats (highlighted as ‘Calc’ in the Appendix), a useful parameter for assessing and monitoring calcareous habitat in Breckland plantations. • Terrestrial snail-associated species-an insect assemblage that can reflect site quality and includes a number of bees that nest in empty snail shells (highlighted as ‘Sn’ in the Appendix). These bees are also calcicoles, though this is not the case with some snail-associated flies and beetles. • Aerial nesting solitary species (excluding cleptoparasites)-to reflect the quantityand varietyofdeadwood and otheropportunities foraerial nesters. • Ground-nesting solitary species (excluding cleptoparasites) to reflect the quantity, quality and variety ofground-nesting conditions. • Coastal-biased species-another useful parameter for Breckland sites which support unusually high concentrations of species most typically associated with coastal dunes (species highlighted as ‘Coastal’ in the Appendix). • Conifer-associated species-a useful parameter for some other insect groups (e.g. flies), though only one such species exists within the \\^i-Passaloecus eremita Kohl (highlighted as Conif in the Appendix). • Cleptoparasites (ofboth solitary and social species)-which can beconverted into a ‘cleptoparasitic load’ by turning the number ofcleptoparasites into a percentage ofthe total fauna. Results Species richness Large, representative samples ofbees and wasps were obtained from all four sites, particularly Elveden, Longleat and Sherwood which were subject to a longer period ofstudy than Whinfell. The data obtained were used to evaluate the quality of the assemblages present ateach site. The various scores that underpin thisevaluation are presented in Table 1. All the species recorded, and the year(s) are listed in the Appendix, which also furnishes briefnotes for the scarcer species, plus any points of interest relating to more frequent ones. A total of 259 species (excluding Apis melliferci L.) was recorded from the four sites, a figure that represents over halfofthe British bee and wasp fauna and exceeds some county lists. The richest site by far was Elveden, with a total of203 confirmed species and one unconfirmed one {Dryudella pinguis (Dahlbom)-suggested by the presence ofits cleptoparasite) based on 11 year’s recording. Longleat produced a list of 136 confirmed species plus one unconfirmed one {Melitta leporina (Panzer) suggested by its cleptoparasite) based on ten year’s recording. Sherwood produced 124 species based on ten year’s recording and Whinfell produced 66 species based on only three year’s recording. The Elveden list is one of the best modern site lists in Britain-there are probably only a handful of sites nationally with modern lists exceeding 200 species. Nine species were added to the Elveden list in the final year of recording, suggesting that this list remains incomplete. Elveden also produced two exceptional single day lists of96 species on 28 July 2002 and 97 species on 19 July 2004. Whilst the lists for the other three sites are somewhat shorter then that for Elveden, they still appear to be very good within their respective counties. 28 BR.J. ENT.NAT. HIST.,20:2007 Table 1. Hymenoptera quality scores from the four British Centerparc villages, using a variety ofparameters. Rarity scores were adjusted for obvious misgradings ofspecies (see Appendix^ Longleat Elveden Sherwood Whinfell Confirmed species (unconfirmed) 136 (137) 203(204) 124 66 Nationally Endangered (RDBl) 0 0 0 0 Nationally Vulnerable (RDB2) 0 2 0 0 Nationally Rare (RDB3) 2 6 0 0 Nationally Scarce (N) 8 34 6 3 Rarity Score (based on RDB & N spp) 600 2500 300 150 Species Quality Index 4.41 12.32 2.42 2.28 Aerial-nesting solitary species (%) 29 (21.3) 49 (24.1) 29 (22.6) 12 (18.2) Ground-nesting solitary Species 42 69 55 17 Calcicolous species 3 6 0 0 Snail-nesting species 2 2 0 0 Acidophiles 6 6 10 6 Coastal-biased species 0 7 0 0 Conifer-associated species 0 0 1 0 Cleptoparasites (load) 30 (22.1 ) 45 (22.2 ) 26 (20.1 ) 19 (28.8 ) The presence ofrare and scarce species Some very significant records were obtained, including several species additional to the recently published list for Cumbria (Robertson, 2005). These scarcer species are highlighted in the Appendix. All sites produced records ofrare or scarce species, though the number ofthese and the resultant SQI value varied greatly (see Table 1). Elveden produced a rarity score of2500 derived from the presence ofeight Red Data Book species and 34 Nationally Scarce ones. Longleat scored 600due to the presence of two Red Data Book species and eight Nationally Scarce ones. Sherwood scored 300 due to the presence of six Nationally Scarce species and the under-recorded Whinfell scored 150 due to the presence of three Nationally Scarce species. Converting the scores to a species quality index gave the same ranking, though this calculation has been found by the author to create quite distorted impressions of relative site quality ona numberofoccasions. In this study, the SQI values suggested that Sherwood and Whinfell were ofsimilar quality. But this is quite unlikely given the northern location and less sandy nature of Whinfell and would probably be conhrmed with more surveying ofWhinfell. The quality ofsome key species assemblages Seventy-one aerial nesters (excluding cleptoparasites), seven calcicoles, two snail- users, 15 acidophiles and sevencoastal-biased specieswere recorded. The numbers of these found at each site are summarised in Table 1. Elveden had the highest number of aerial nesters (49) and the highest proportion of aerial nesters within its fauna, though the latter figure was only slightly greater than the figures for Sherwood and Elveden. Elveden also had the greatest number ofcalcicoles (6) with Longleat being the only other site capable of supporting any (3). Both sites supported two snail- users, which are both strongly calcicolous {Osmia hicolor (Schrank) and Hoplitis spinulosa (Kirby)). Only Elveden was capable of supporting any coastal-biased species, an assemblage of seven. Sherwood produced the highest number of acidophilous species (10), with the other three sites each supporting six. Sherwood BR.J.ENT.NAT. HIST.,20:2007 29 was the only site found to be supporting the conifer-requiring Passaioecus eremita. Thenumberofcleptoparasiteswasgreatestat Elveden, thoughasaproportion ofthe total aculeate fauna, the load’ was not significantly greater than that for either Longleat or Sherwood. By comparison, the meagre fauna of 19 cleptoparasites at Whinfell actually constituted a load ofnearly 30% ofthe entire fauna, a value much higher than the other three sites. Discussion The rarity scores forElveden, Longleatand Sherwood indicate thatthey should be considered important sites for aculeate Hymenoptera. Rarity scores for some other key sites recently subject to comparable surveys, based on modern lists ofRed Data Book and Nationally Scarce species, include 1200 for Highgate Common, Staffordshire (a heathland supporting the richest bee and wasp fauna currently known in the West Midlands) and 650 for the Bishops HilbBishops Bowl complex, Warwickshire (a limestone site supporting the second richest bee and wasp fauna currently known in the West Midlands) (Falk, 2006). The rarity score for Elveden should be viewed as reflecting a bee and wasp assemblage ofnational significance. Factors influencing species richness A numberoffactors appeared to influence thedifferent species richness ofthe four sites: (i) Theextentandqualityoffloristicallydiversehabitat. Elvedencurrently supports considerably more semi-natural grassland than the other three Centerparcs. What is more, this grassland has formed on calcareous chalky sands and has a character resembling that of other classic Breckland sites such as Wangford Warren, Red Lodge and East Wretham Common. These grasslands are characterised by very rich floras containing both calcicoles and typical heathland plants and with high densities of flowers throughout summer months that attract a rich fauna. The abundance of umbellifers such as hogweed and parsnip was found to be critical. Parsnip, in particular, plays an important role as a forage plant for bees and wasps at Breckland sites, and it was largely the persistent surveying ofsome very large stands at Elveden that produced the two exceptional single day lists noted above. Hogweed was present at parts ofWhinfell and was ofmajor importance there, but Longleat and Sherwood supported very few umbellifers, which almost certainly reduced the potential ofthese sites for bees and wasps. As a rule ofthumb, acidic soils tend to be less floristically diverse and produce a less continuous sequence of useful flowers compared with calcareous sites, though the heather flowers that can be present in late summer are very important. Spring blossoming shrubs such as Sa/ix, Primus and Ulex are very important in March and April for Andrena and Nomada bees. But whilst these shrubs occurred widely at Elveden and Sherwood, they were scarce at Longleat and Whinfell, which was reflected in their much smaller spring bee faunas. But aside from the variety offlowers present, the sheer extent of open habitat within these coniferous settings was probably crucial as many bee and wasp species cannot survive in small patches ofotherwise suitable habitat. This is borne out by the author’s work at various Warwickshire plantations that are not diversified or only slightly so. Simple ride systems and small clearings tend not to support rich aculeate faunas (e.g. 30 BR.J.ENT.NAT. HIST.,20:2007 Hay Wood SP2071 and Birchley Hays Wood SP2684), but the fauna will increase where wider rides are created and larger, more permanent patches of grassland are present (e.g. Brandon Wood SP3876 and Oversley Wood SP1056). (ii) Geographic location. Most bees and wasps have a southern bias (often a south- eastern bias), which means that one tends to find smaller assemblages of aculeates furthernorth. It is not a strict rule, as many othervariablescan affect the fauna ofa site and result in some relatively rich sites in the north and poor sites in the south. The best sites in the Brecks and southern heathland districts support over 200 species, whilst the best ones in the Midlands support about 130 species (S. Falk data) and few sites in northern England have lists exceeding 100 species (M. Archerdata). The influence ofgeographical location isclearly reflected in theCenterparc data, even ifthe morecalcareous nature of Elveden and the under-recording ofWhinfell are allowed for. But it should be noted that Whinfell supported three northern-biased species {Bomhus numticola Smith, Crossocerus leucostomus (L.) and Passaloecus mouilicornis Dahlbom) that would not be anticipated from the other three sites. (iii) The surrounding countryside. All four Centerparcs were located within much larger tracts of conifer plantation and other habitats of varying ecological character and quality. Elveden was most fortunate in its location, lying adjacent to the 570ha Lakenheath Warren SSSI, with Wangford Warren and several other important chalk-heath sites within a few miles. This almost certainly facilitated the colonisation ofnew habitats created at this Centerparc over the past 15 years and allowed colonisation by some surprisingly scarce species. It is best to view Breckland bee and wasp assemblages as being comprised of numerous local populations (metapopulations) that can move around the Breckland landscape with different levels of efficiency, taking advantage of any new suitable habitat that forms reasonably close to an existing colony ofa given species. This would help to explain why species such as Andrena haftorfiana and Melitta hciemorrhoidalis (F.) were recorded at Elveden, even though only small quantities oftheir forage plants were present (less than would be needed for viable, permanent populations). It is possibly also the reason why so many species were only recorded once despite careful surveyingovermanyyears. Longleat islocated withinanarea thatlostmuch of its former heathland through coniferisation. The new heathland created is currently very isolated, and the bee and wasp fauna it supports is unexceptional in a southern England context, presumably because there are few nearby sources forsuch species. Theheathland here notablylacked the two heather specialist bees found at Sherwood and Elveden, sites which both occur within districtscharacterised by much heathland. Sherwood’s location within a district supporting particularly extensive areas of lowland heathland may explain why its new heathland supported the highest number of acidophiles, including two heather-associated bees, and why more ground nesting solitary species were present than Longleat (55 vs. 42). (iv) The availability ofdead wood and other aerial nesting opportunities. Whilst the majority of bees and wasps present were ground-nesters, aerial nesters still comprised a significant proportion ofthe fauna. The number ofaerial nesting species was partially dictated by the quantity and variety of dead wood and hollow stems, though the previous three factors also exerted strong influences. The high number of aerial nesters at Elveden in part reflected the greater amount ofdead wood (both large log piles and some standing dead trees), key