R E S E A R C H A R T I C L E

Grassland Restoration on Landfill Sites in the EastMidlands, United Kingdom: An Evaluation of FloralResources and Pollinating InsectsSam Tarrant,1,2,3 Jeff Ollerton,1 Md Lutfor Rahman,1 Joanna Tarrant,1 and Duncan McCollin1

Abstract

Pollinators are declining in Europe due to intensification ofagriculture, habitat loss and fragmentation. Restored land-fill sites are a significant potential reserve of semi-naturalhabitat, so their conservation value for supporting popula-tions of pollinating insects was here examined by assessingwhether the plant and pollinator assemblages of restoredlandfill sites are comparable to reference sites of exist-ing wildlife value. Floral characteristics of the vegetationand the species richness and abundance of flower-visitinginsect assemblages were compared between nine pairs ofrestored landfill sites and reference sites in the East Mid-lands of the United Kingdom, using standardized methodsover two field seasons. No differences were found betweenthe restored landfill and reference sites in terms of speciesrichness or abundance of plants in flower and both types

of site had similar assemblages of pollinators. However,plant and insect assemblages differed across the season,with species richness and abundance being lower for therestored landfill sites in the spring and higher in theautumn compared to the reference sites. The results indi-cate that in this region, landfill sites are being restored to astate comparable to that of the reference sites with regardsto their provision of floral resources and the associatedinsect pollinator assemblages. Since there are currently2,200 working landfill sites in England and Wales, cover-ing 28,000 ha, and closing at a rate of 100 per year, thisis potentially a significant reserve of land that could berestored.

Key words: biodiversity, brown-field, flower-visiting,habitat, pollinator, restored.

Introduction

Populations of some plants and pollinators are decliningin Europe due to habitat degradation, fragmentation, andagricultural intensification (Allen-Wardell et al. 1998; Corbet2000; Steffan-Dewenter et al. 2002). For example, in theUnited Kingdom intensification of agriculture has led to a 97%loss of unimproved grasslands in England and Wales (Dryden1997), and many pollinating insect species are experiencingsteep declines, both in population size and range (Beismeijeret al. 2006; Potts et al. 2010).

Biotic pollination by insects and other animals is a vitalecological process in almost all terrestrial ecosystems, and itis estimated that globally over 87% of angiosperms requireanimal pollinators (Ollerton et al. 2011). Pollinators maintainpopulations of wild plants and therefore, indirectly, the

1Landscape and Biodiversity Research Group, School of Science and Technology,The University of Northampton, Avenue Campus, Northampton, NN2 6JD, U.K.2Present address: RSPB UK Headquarters, The Lodge, Sandy, Bedfordshire SG192DL, U.K.3Address correspondence to S. Tarrant, email sam.tarrant@rspb.org.uk

© 2012 Society for Ecological Restorationdoi: 10.1111/j.1526-100X.2012.00942.x

ecosystem goods and services that human society relies upon.In addition, many crop plants rely on the direct ecosystemservices provided by wild pollinators, estimated to be worth¤153 billion annually, equivalent to 9.5% of the value ofglobal agricultural output (Gallai et al. 2009). In the UnitedKingdom, the value of insect pollination is approximately£430 m p.a., some 13% of farming income (UK NationalEcosystem Assessment 2011), and it has recently been sug-gested that most of this service is delivered by wild pollinatorsrather than managed honey bees (Breeze et al. 2011).

One partial solution to halting the decline of pollinatorsin the United Kingdom may be the creation or restoration ofsuitable habitat. For example, landfill sites are often restoredas grassland for stock grazing or simply as an expanse of semi-natural vegetation (Gilbert & Anderson 1998; Watson & Hack2000). There are approximately 2,200 working landfill sites inEngland and Wales, covering 28,000 ha and they are closingat a rate of about 100 per year (Environment Agency forEngland and Wales 2006). Following closure, landfill sites areusually capped with clay and then covered with approximately1 m of soil before being hydro-seeded with a liquid emulsionof grass seed, fertilizer, and binding agent, or simply left toregenerate naturally (Watson & Hack 2000; Hutchings et al.2006).

560 Restoration Ecology Vol. 21, No. 5, pp. 560–568 SEPTEMBER 2013

Flowers and Pollinating Insects on Restored Landfills

Unimproved flower-rich grassland has been found to have ahigh abundance and richness of pollinating insects (Ockinger& Smith 2007). However, we know virtually nothing aboutthe conservation value of restored landfill sites for pollinatinginsects, though other restored habitats are known to bevaluable, for example, reclaimed opencast mines in NorthAmerica (Holl 1995).

The aim of this research was to establish the potentialof restored landfill sites to support assemblages of insectpollinators. We compared the species richness, abundance, andassemblage composition of insect pollinated flowering plants(and by implication, the floral resources they provide) betweenpairs of restored landfill sites and reference sites of knownnature conservation value. Comparisons were also madebetween those restored landfill sites which had been artificiallysown and those which had naturally revegetated. At the sametime, we compared the abundance, richness, and compositionof flower-visiting insect pollinators to those plants.

Methods

Study Sites

Nine restored landfill sites were selected in the East Midlands,United Kingdom (Table 1). The sites had similar characteristicsand were representative of restored landfill sites within theregion and probably within lowland Britain as a whole. Theselection criteria for the landfill sites were greater than or equalto 50% of the site restored (to avoid undue influence fromongoing landfilling operations), greater than or equal to 0.5 hain area and restored for greater than or equal to 4 years toallow establishment of vegetation.

Comparison reference sites were the closest grassland sitesof recognized nature conservation value, being designatedas either Local Nature Reserves (LNRs) or Sites of SpecialScientific Interest (SSSI) (Table 1). The rationale for selectingnearby designated conservation areas was that such sitesprovide a potential “end-point” or “target” in terms ofecological restoration (Handel et al. 1994). The reference sitesexperienced similar local climates, had the same regional plantand insect species pools, and comparable landscape contexts(mean distance = 4.5 ± 3.5 km, range = 1.3–11.8 km). Theselection criteria for reference sites were ≥0.5 ha in area andnot undergoing any specific pollinating insect conservationmeasures. All nine pairs of sites were surveyed in the firstyear of the study (2007). In the following year (2008), thiswas reduced to three pairs of sites in order to increase theintensity of sampling.

Survey Methods

Surveys were conducted from March to October in 2007 and2008; this corresponds to the main flowering period in centralEngland and hence to flower-visitor activity (Fitter et al. 1995).All sites were surveyed three times each during the fieldworkseason, in Spring, Summer, and Autumn. Paired sites were

sampled on consecutive days whenever weather conditionspermitted to reduce temporal bias.

Standardized plant surveys were used (Dicks et al. 2002;Potts et al. 2006). Transects (100 × 2 m) were centeredfrom the approximate middle of the site and orientated usingrandomized bearing tables. All flowering plants were identifiedto species level using Stace (1991) and Kent (1992), and thetotal number of flowering units recorded. One floral unit wasdefined as a head (e.g. Trifolium pratense), an umbel (e.g.Daucus carota), or a capitulum (e.g. Centaurea scabiosa).“Flowering plants” henceforth refer to insect pollinated plantswhich were in flower during these surveys.

A “floral cover” method to represent available floralresources was used which combines floral abundance withinflorescence size. Mean area of the floral unit from abovewas measured for each flowering plant species and then mul-tiplied by their frequencies. The sizes of the floral units weredetermined through field sampling, and checked for consis-tency using previously published literature. Only open flowerslikely to produce nectar or pollen were recorded. The restoredlandfill sites were assessed for their average floral cover (meancm2 floral cover per m2 habitat) and their total area of floralcover. Floral cover is sufficiently robust to allow comparisonsacross sites and has been used in previous studies (Steffan-Dewenter & Tscharntke 2001; Potts et al. 2006; Meyer et al.2009). Flower-visitor surveys were undertaken three timesbetween 09:00 and 16:00 hours on days which were warm andsunny with little or no wind, as outlined in the Butterfly Mon-itoring Scheme (Pollard & Yates 1993). Surveys each lasted30 minutes and all flower-visiting insects seen to be feedinglegitimately (i.e. not nectar robbing) and large enough to touchanthers and stigmas were captured.

In the first year of study, plants in flower and flower vis-itors were surveyed using the same transects as for the flo-ral resources surveys. The transect was left undisturbed for20 minutes following the initial plant survey to allow theflower visitors to return. Each transect was surveyed at a rate ofapproximately 3 m/minute for 30 minutes. All insects observedto touch the sexual parts of flowers were either captured usinga butterfly net and transferred into individually labeled spec-imen jars, or directly captured into the jars. After the surveywas completed, those insects that could be identified in thefield were recorded and released. The flower-visitor surveyswere conducted in the morning, within 1 hour of midday, andin the afternoon to sample those insects active at differenttimes. Insects that could not be identified in the field werecollected as voucher specimens for later identification. Identi-fications were verified using reference collections and by taxonspecialists.

Relatively low capture rates in the first year led to methodsbeing altered in the second year when surveying followed aspiral pattern from a randomly determined point on the sites,at a standard pace of 10 m/minute for 30 minutes, followingNielsen and Bascompte (2007) and Kalikhman (2007). Givena 2-m wide transect, an area of approximately 600 m2 wassampled in each survey.

SEPTEMBER 2013 Restoration Ecology 561

Flowers and Pollinating Insects on Restored Landfills

Table 1. The restored landfill sites and reference sites.

Site Type Status Operator Latitude Longitude Age Size (ha) Revegetation Pairing Date

Bletchley Landfill site Working WRG 51◦58′57′′N 0◦45′26′′W 5 34.00 Sown 1 2007Brixworth Closed Sita 52◦20′32′′N 0◦53′26′′W 8 11.25 Natural 2 2007/2008Brogborough Working WRG 52◦02′44′′N 0◦35′31′′W 5 26.11 Sown 3 2007Cranford Working Sita 52◦22′31′′N 0◦37′50′′W 4 0.58 Natural 4 2007Harlestone Working BPL 52◦16′01′′N 0◦57′53′′W 4 6.60 Natural 5 2007Kettering Closed BPL 52◦25′34′′N 0◦43′09′′W 15 10.80 Natural 6 2007Kilsby Working Biffa 52◦19′00′′N 1◦10′07′′W 12 7.19 Natural 7 2007Sidegate Lane Working Sita 52◦19′35′′N 0◦39′23′′W 15 4.13 Sown 8 2007/2008Wootton Closed Viridor 52◦11′26′′N 0◦53′28′′W 7 14.57 Sown 9 2007/2008Barnes Meadow Reference site LNR WT 52◦13′44′′N 0◦52′23′′W — 4.18 — 9 2007/2008Blue Lagoon LNR MKBC 51◦59′20′′N 0◦44′19′′W — 1.05 — 1 2007Ditchford LNR WT 52◦18′03′′N 0◦38′10′′W — 12.19 — 8 2007/2008Draycote SSSI WWT 52◦20′00′′N 1◦20′20′′W — 2.25 — 7 2007Glebe Meadow LNR BCC 52◦03′17′′N 0◦28′34′′W — 4.12 — 3 2007/2008Pitsford SSSI WT 52◦19′38′′N 0◦51′32′′W — 0.79 — 2 2007River Ise Meadows SSSI WT 52◦26′11′′N 0◦44′17′′W — 21.36 — 6 2007Scrub Field LNR WT 52◦16′06′′N 0◦52′48′′W — 3.00 — 5 2007Twywell LNR WT 52◦23′09′′N 0◦37′11′′W — 8.08 — 4 2007

Status: Working, landfills partly in use; Closed, landfills wholly restored; LNR, Local Nature Reserve; SSSI, Site of Special Scientific Interest. Operator: WRG, Waste RecyclingGroup; Sita, Sita UK; BPL, Barton Plant Limited; Biffa, Biffa Waste Services; Viridor, Viridor Waste Management; WT, The Wildlife Trust for Bedfordshire, Cambridgeshire,Northamptonshire, and Peterborough; MKBC, Milton Keynes Borough Council; WWT, Warwickshire Wildlife Trust; BCC, Bedfordshire County Council.Age refers to years after restoration of the landfill, up to 2009.

Table 2. The most abundant plant species on the restored landfill sites and reference sites.

2007 2008

Site type Plant speciesFloral abundance(units per 200 m2)

Frequencyon sites Plant species

Floral abundance(units per 200 m2)

Frequencyon sites

Restored landfill Trifolium dubium 337.6 6 of 9 Trifolium repens 120.5 3 of 3Picris echioides 153.8 8 of 9 Lotus corniculatus 98.6 1 of 3Trifolium repens 133.1 6 of 9 Cirsium arvense 60.4 2 of 3Picris hieracioides 67.6 8 of 9 Trifolium dubium 52.8 3 of 3Cardamine flexuosa 66.2 2 of 9 Ranunculus repens 32.7 3 of 3

Reference Trifolium dubium 483.3 7 of 9 Ranunculus acris 376.5 3 of 3Ranunculus bulbosus 162.0 5 of 9 Lotus corniculatus 253.5 2 of 3Lotus corniculatus 148.2 7 of 9 Stellaria graminea 135.6 1 of 3Ranunculus acris 138.5 8 of 9 Galium verum 121.4 1 of 3Galium verum 135.5 2 of 9 Rhinanthus minor 114.2 1 of 3

Floral abundance refers to the mean number of floral units per 200 m2 survey of a particular plant species, across all sites.

Data Analysis

Data were tested for normality using one-sampleKolmogorov–Smirnov tests. For testing differences in meansor medians, paired t-tests, one-way analysis of variances(ANOVAs), Mann–Whitney U -tests and Kruskal–Wallistests were used. Pearson’s correlations were used to testassociations between variables.

Insect pollinated flowering plant species composition andfloral abundance between sites by type were representedby non-metric multidimensional scaling (NMDS). Euclideandistance was used rather than the Bray-Curtis measure as itgives a greater distortion and hence visual spread to the data,which is an advantage when the sites being compared areexpected to be very similar (Kessel & Whittaker 1976). Unlikeother ordination techniques NMDS does not make assumptionsabout the distribution of the variables (Maarel 2005) and

uses instead rank distance for ordination with similar sitesbeing closer together in the plot (Legendre & Legendre 1998).This method is sensitive to showing outliers and the distancebetween points shows the relative similarity (McCune & Grace2002; Ollerton et al. 2009).

Results

Plant Assemblages and Floral Resources

Over the duration of this study, 125 floral transects weresurveyed encompassing 25,000 m2 and approximately 138,000floral units were counted from 98 plant species. On the restoredlandfill sites a total of 63 species of flowering plant were found,with 19 exclusively on the restored sites. On the reference sites74 species were found, 30 exclusively. The most abundant

562 Restoration Ecology SEPTEMBER 2013

Flowers and Pollinating Insects on Restored Landfills

Figure 1. NMDS ordination of plant species and abundance for sownand naturally revegetated restored landfill site (1–9) and reference sites(A–I) for 2007. Restored landfill sites: 1, Bletchley; 2, Brixworth; 3,Brogborough; 4, Cranford; 5, Harlestone; 6, Kettering; 7, Kilsby; 8,Sidegate Lane; 9, Wootton (NB for clarity; 5, Harlestone is natural; 9,Wootton is sown). Reference sites: A, Barnes Meadow; B, Blue Lagoon;C, Ditchford; D, Draycote; E, Glebe Meadow; F, Pitsford; G, River IseMeadows; H, Scrub Field; I, Twywell. Two-dimensional S-stress = 0.028.A NMDS S-stress value below 0.1 shows that the data is representedfaithfully (McCune & Grace 2002).

plant species in restored and corresponding reference sites areshown in Table 2. NMDS analysis showed that the majorityof sites were closely clustered (Fig. 1), with only three ofeach type of site as outliers, probably due to the abundance ofidiosyncratic species that occurred on single sites.

The restored landfill sites and reference sites were similarin their mean annual richness of plants in flower (2007:t = −0.94, df = 8, p = 0.37; 2008: t = 0.20, df = 2, p = 0.86). Attheir flowering peak in 2007, the nine restored landfill sites(mean size = 12.8 ha) had a mean floral cover of 6.6 cm2/m2

and a combined total floral cover of 643 m2. This compareswith the nine reference sites (mean size = 6.3 ha), having amean floral cover of 10.1 cm2/m2 and a combined total floralcover of 342 m2. In both cases, the mean annual totals were thesum total of the mean seasonal values. However, there wereseasonal differences in the richness of plants in flower in both2007 and 2008 (Fig. 2), with reference sites accumulating noadditional floral cover in the autumn period, in contrast to therestored landfill sites (Fig. 3).

Flower-Visiting Insects

A total of 201 flower-visitor surveys were performed, 129 inthe first year on 18 sites and 72 in the second year on sixsites. Over the two field seasons, 942 flower-visiting insectswere sampled. In 2007, the restored landfill sites yielded 156individuals from 30 species of flower-visiting insects, and thereference sites 161 individuals of 37 species. In 2008, therestored landfill sites yielded 405 individuals of 41 insectspecies and the reference sites 220 individuals of 40 species.

The distribution of species between taxonomic groups isshown in Table 3. The restored landfill and reference siteshosted similar numbers of most groups of flower visitors in

(a)

(b)

Figure 2. Average seasonal species richness of plants in flower forrestored landfill sites and reference sites (mean ± 95% CI, N = samplesize). (a) 2007 ANOVA—Landfill sites across seasons F [2,16] = 6.40p = 0.01; Reference sites across seasons F [2,14] = 5.68 p = 0.02. Pairedt-tests (two-tailed) Spring Landfill versus Reference t = −0.34 df = 3p = 0.75, Summer Landfill versus Reference t = 0.45 df = 5 p = 0.68. (b)2008 ANOVA—Landfill sites across seasons: F [2,8] = 2.76 p = 0.14,Reference sites across seasons: F [2,8] = 11.04 p = 0.01. Paired t-tests(two-tailed); Spring Landfill versus Reference: t = −1.51 df = 2 p = 0.27,Summer Landfill versus Reference t = 0.43 df = 2 p = 0.70. Wilcoxon test(two-tailed): Autumn Landfill versus Reference Z = −1.34 p = 0.18.

both years, though variations between site type and betweenyears were apparent.

In both 2007 and 2008, there was no significant differencein mean annual flower-visiting insect species richness per sitein restored landfill compared to reference sites. However, therewere seasonal differences between site types (Fig. 4).

SEPTEMBER 2013 Restoration Ecology 563

Flowers and Pollinating Insects on Restored Landfills

(a)

(b)

Figure 3. Cumulative seasonal floral cover for restored landfill andreference sites in 2007, expressed as: (a) mean seasonal floral cover(cm2/m2) per site and; (b) total on-site mean seasonal floral cover (m2)per site.

The most abundant native pollinators for the restoredlandfill sites in both years were the bumblebees Bombusterrestris/lucorum and Bombus lapidarius , plus various fliesand hoverflies (Table 4), whereas the most abundant specieson the reference sites in both years were the bumblebeesBombus pascuorum , B. terrestris/lucorum , and B. lapidarius ,and the fly Calliopum sp. (Diptera: Lauxaniidae). The non-native honeybee (Apis mellifera) was also abundant in 2007,but not 2008. The restored landfill and reference sites thereforeshared a number of their most abundant species (Table 4).

With the exception of one outlier, Brixworth (site 2), theNMDS ordination of pollinators in 2007 showed a clearseparation between the sites along the vertical axis (Fig. 5).However, eight of the nine restored landfill sites were clusteredtogether within the plot, indicating that they share many oftheir flower-visitor species. The reference sites have greatervariance in their spread, indicating they were less similar toone another. The Scrub Field (H), for example, had the highestabundance of A. mellifera as well as the unique presence ofHelophilus pendulus , and Barnes Meadow (A) had the highestabundance of B. pascuorum.

Discussion

Previous research into plant diversity on landfill sites in theUnited Kingdom during the late 1980s and early 1990s foundrelatively low species richness, attributed to poor restorationstandards and containment measures, where methane emis-sions through the soil affected plant growth (Wong 1988;Ireland 1991). The policies brought in to regulate the restora-tion and after-use of modern landfill sites, coupled with currentrestoration practices, appear to have improved this situationand have had the effect of creating well vegetated, relativelyflorally diverse sites that support diverse assemblages of pol-linating insects.

Floral characteristics were similar between the restoredlandfill and reference sites, with no difference in either plantrichness or abundance of floral resources in 2007. In addition,both the sown and naturally revegetated restored landfill siteswere similar, though further studies need to be undertaken toassess the specific benefits of both methods of revegetation(Gilbert & Anderson 1998; Watson & Hack 2000; Hutchingset al. 2006).

The assemblages of flower-visiting insects that we recordedon these sites had relatively modest diversity because theywere intended to be samples rather than complete surveys.This sampling showed that there were no differences in thespecies richness and abundance of flower-visiting insects onthe restored landfill sites compared to the reference sites,paralleling the findings from the plant surveys. NMDS analysisfor 2007 showed eight of the nine restored landfill sites

Table 3. Species richness for the main taxa of flower-visiting insects recorded on restored landfill sites and reference sites.

2007 2008

Restored landfill Reference site Restored landfill Reference site

Bees (except Bombus) Hymenoptera: Apoidea 2 5 6 2Bumblebees Hymenoptera: Bombus 5 4 4 5Beetles Coleoptera 2 3 3 3Butterflies Lepidoptera 1 6 6 8Flies Diptera 6 8 4 8Hoverflies Diptera: Syrphidae 11 10 16 12Other — 3 1 2 2Total — 30 37 41 40

Bees, all non-bumblebee bees; Bumblebees, Bombus and Psithyrus; Butterflies, all butterflies and day-flying moths; Flies, all non-Syrphidae; Other, all other flower-visitinginsects.

564 Restoration Ecology SEPTEMBER 2013

Flowers and Pollinating Insects on Restored Landfills

(a)

(b)

Figure 4. Average seasonal flower-visitor abundance per survey forlandfill sites and reference sites (mean ± 95% CI, N = sample size). (a)2007 (survey = 30 minutes and 200 m2) ANOVA—Landfill sites acrossseasons F [3,56] = 2.76 p = 0.05; Reference sites across seasonsF [2,40] = 1.80 p = 0.18. Independent t-tests (two-tailed) Spring Landfillversus Reference t = −2.17 df = 18.78 p = 0.04, Summer Landfillversus Reference t = 0.42 df = 31 p = 0.68, Autumn Landfill versusReference t = 1.93 df = 19 p = 0.07. (b) 2008 (survey = 30 minutes and600 m2) ANOVA—Landfill sites across seasons F [2,41] = 5.74 p = 0.01;Reference sites across seasons F [2,35] = 14.60 p < 0.01. Independentt-tests (two-tailed) Spring Landfill versus Reference t =−2.64 df = 25p = 0.02; Summer Landfill versus Reference t = 1.83 df = 19.62p = 0.08. Mann–Whitney two-independent samples test Autumn Landfillversus Reference U = 15.00 p = 0.01.

were placed close together within the plot, indicating theyshared many flower-visitor species, whereas the reference siteshad greater variance in their spread, indicating they wereless similar to one another. There is clearly a large overlapbetween the restored landfill and reference sites in terms ofthe composition of the flower-visiting insect assemblages.

There were significant differences between the site types intheir seasonal plant species richness and abundance of floralcover, with the reference sites having greater richness andabundance in the spring, and the restored landfill sites greaterrichness and abundance in the autumn. In 2007, the referencesites had a greater mean cumulative seasonal floral cover perm2 than the restored landfill sites. In 2008, this was reversed,although the larger size of the restored landfill sites meant thaton a landscape scale they were providing nearly twice as manyfloral resources as the reference sites.

Across the seasons, the restored landfill sites were found tohave lower flower-visitor richness and abundance in the spring,but greater in the autumn, compared to the reference sites. Thisis probably due to the reduced floral resources available onthe reference sites in the autumn; all of the reference sites aremown in late summer, reducing the abundance of flowers. Therestored landfill sites are therefore providing a valuable reserveof floral resources for those flower-visiting insects active laterin the season. The decreased spring insect abundance onthe restored sites is probably due to the lower mean floralcover per m2 in the spring as previous studies have shown apositive correlation between flower and pollinator abundancein temperate grassland (Hegland & Boeke 2006). However,it is also possible that a lack of suitable over-wintering sitescould be a contributory factor limiting insect abundance; manyflower-visiting species require suitable places to hibernate orto remain dormant until the spring (Alford 1969; Goulson2003), although there is a lack of available information onthis (Carvell et al. 2004). If such habitats are not present,then the restored landfill sites may be de-colonized eachwinter and require time to be re-colonized in the spring.Alternatively, the restored landfill sites may be providing floralfood resources but the flower-visiting insects may be nestingand hibernating elsewhere. This may mean that the residentflower-visiting insects are utilizing the surrounding landscapefor food resources, then switching to the restored landfill sitesonce floral resources have built up (Osborne et al. 2008).

The findings for this study are similar to the results fromthe (albeit limited) previous research studying plant-pollinatorassemblages in British habitats following restoration. Onrestored versus established hay meadows, for example, speciesrichness and abundance were the same, though few flower-visiting insect species were shared among the sites (Forup& Memmott 2005). On four pairs of restored and ancientheathlands, the restored sites had fewer insect species thantheir paired sites in 2001, but more species in 2004 (Forupet al. 2008). However, it should be noted that in both of thesestudies, the restored sites were directly targeted at recreatingthe same vegetation structure as that of the reference sites,which is not the case in this study.

SEPTEMBER 2013 Restoration Ecology 565

Flowers and Pollinating Insects on Restored Landfills

Table 4. Most abundant flower-visiting insect species on restored landfill sites and reference sites.

2007 2008

Site type Insect speciesAbundance

(total individuals)Frequency

on sites Insect speciesAbundance

(total individuals)Frequency

on sites

Restored Apis mellifera 34 6 of 9 Calliopum sp. 77 2 of 3landfill Calliopum sp. 19 6 of 9 Bombus terrestris/lucorum 49 3 of 3

Bombus terrestris/lucorum 12 6 of 9 Bombus lapidarius 43 3 of 3Bombus pascuorum 10 5 of 9 Episyrphus balteatus 33 2 of 3Bombus lapidarius 10 4 of 9 Oedemera nobilis 27 3 of 3

Reference Apis mellifera 33 5 of 9 Bombus lapidarius 26 3 of 3Bombus pascuorum 26 6 of 9 Bombus pascuorum 25 3 of 3Calliopum sp. 17 5 of 9 Calliopum sp. 24 3 of 3Bombus lapidarius 15 8 of 9 Bombus terrestris/lucorum 17 3 of 3Bombus terrestris/lucorum 6 4 of 9 Lasioglossum sp 16 3 of 3

Abundance, total number of individuals recorded for all sites of that type; Site frequency, the number of sites on which that species was present.The species B. terrestris and B. lucorum were counted as a single taxon due to the difficulty of separating them in the field.

Figure 5. NMDS ordination of flower-visitor species and abundance onrestored landfill sites (1–9) and reference sites (A–I) for 2007.Two-dimensions used, S-stress = 0.142. Restored landfill sites: 1,Bletchley; 2, Brixworth; 3, Brogborough; 4, Cranford; 5, Harlestone; 6,Kettering; 7, Kilsby; 8, Sidegate Lane; 9, Wootton. Reference sites: A,Barnes Meadow; B, Blue Lagoon; C, Ditchford; D, Draycote; E, GlebeMeadow; F, Pitsford; G, River Ise Meadows; H, Scrub Field; I,Twywell. S-stress = 0.14.

Dixon (2009) has recently argued that when restoringhabitats, their potential to support flower-visiting insects, andthe subsequent pollination of flowering plants, are importantconsiderations. Restoration may be much more successful ifdue attention is applied to pollinators in the early stages (Neal1998). The direct effects of pollination within a restored habitatmay be quite apparent, e.g. seed set. The indirect effects, forexample, gene flow and genetic diversity within populations,may be less apparent, but are nonetheless essential for thelong-term success of restoration.

Although these findings suggest a positive role for restoredlandfill sites in supporting pollinating insects within lowlandEngland, the presence of common and abundant speciesshared between types of sites may be indicative of the homog-enization of assemblages of pollinators in this region. Thosespecies of insect which are common and able to adapt to theintensive agricultural landscape may be increasingly more

common throughout the rural landscape. Conversely, thosespecies which are rare and doing poorly within the landscapemay be becoming rarer or extinct. This would parallel thebiotic homogenization of the British landscape more broadly,for example, in vascular plant species (McCollin et al. 2000;Keith et al. 2009). This clearly has significant implicationsfor the conservation of biodiversity. Such patterns can easilybe missed when relatively small scale studies are looked atin isolation and more work is required to examine speciespresence and abundance in habitats across the landscape andregionally.

Despite these caveats, restored landfill sites have thepotential to play a role in reversing declining populations offlower-visiting insects by at least partly replacing importantpollinator habitats such as unimproved flower-rich grasslands.This may have benefits in relation to the ecosystem servicesprovided by pollinators in the United Kingdom. It should benoted that few natural areas are managed or valued for theecosystem services they provide, although many are managedfor the goods they produce. This may be because the ecol-ogy of ecosystem services such as crop pollination is poorlyknown, limiting our ability to understand their value. Publicperception is that bees (and particularly honeybees) are ourprimary pollinating insects. However, relying solely on honeybees will not be as successful as an assemblage of nativeinsect pollinators (Goulson 2003; Winfree et al. 2007). Nativeunmanaged bee populations provide important pollination ser-vices to various crops (Kremen et al. 2004), and are generallymore diverse and abundant near to natural habitat (Dormannet al. 2007; Klein et al. 2007). One suggested agriculturaloption is to provide semi-natural habitat for native pollina-tors since their pollination services produce higher crop yieldsbut they require relatively little land (Ricketts et al. 2008).Restored landfill sites could help to achieve that goal, partic-ularly if they are connected to fields via hedgerows that canserve as conduits to pollinating bumblebees (Cranmer et al.2011). In conclusion, this research has shown that restoredlandfill sites can support a significant diversity of plant andpollinating insect species. The conservation potential of the

566 Restoration Ecology SEPTEMBER 2013

Flowers and Pollinating Insects on Restored Landfills

2,200 landfill sites (28,000 ha) is comparable to the sum totalof all LNRs in terms of distribution and land area: thereare more than 1400 reserves covering 35,000 ha in England(Natural England 2009). This promises a practical solutionfor the conservation of populations of flower-visiting insects.Restored landfill sites can operate as long-term wildlife habi-tat areas within the wider agricultural environment. Restoredlandfill sites should be managed to maximize floral resourceprovision in the autumn, when they may be absent within thelandscape due to late summer grazing and mowing. Habitatsthat provide a seasonal succession of suitable forage plantscan support larger and more diverse pollinator assemblages(Fussell & Corbet 1992; Corbet 1995). The maintenance ofhealthy and diverse populations of pollinating insects in therural environment may ensure maximum yields of agriculturalplants (Kremen et al. 2004) and will also be of great value forthe survival and propagation of native plant species (Osborne& Williams 1996).

Implications for Practice

• The restoration of landfill sites as sown or naturallyrevegetated grasslands has been shown to support pop-ulations of pollinators by providing them with floralresources, particularly later in the season when they maybe absent elsewhere. Restored landfill sites should bemanaged to provide floral resources earlier in the seasonto benefit flower-visiting insects, by paying attention tomowing regimes and also sowing suitable early floweringplants.

• Vegetation should be managed as a mosaic in whichsome is left standing until spring to provide over-wintering habitat for some pollinators.

Acknowledgments

This research was part of a PhD project funded by theSITA Environmental Trust with money from the U.K. LandfillTax Credit Scheme. We are grateful to the various landfilloperators, local councils, and Wildlife Trusts for giving uspermission to undertake fieldwork on their sites. We aregrateful to Dr. Margaret Bates from the Centre for SustainableWastes Management, University of Northampton, for valuablesuggestions during this project. Finally, thanks to the reviewerswhose suggestions improved the manuscript.

LITERATURE CITEDAlford, D. V. 1969. A study of the hibernation of bumblebees (Hymenoptera:

Bombidae) in Southern England. Journal of Animal Ecology 38:149–170.

Allen-wardell, G., P. Bernhardt, R. Bitner, A. Burquez, S. Buchmann, J.Cane, et al. 1998. The potential consequences of pollinator declineson the conservation of biodiversity and stability of food crop yields.Conservation Biology 12:8–17.

Beismeijer, J., S. P. M. Roberts, M. Reemer, R. Ohlemuller, M. Edwards, T.Peeters, et al. 2006. Parallel declines in pollinators and insect-pollinatedplants in Britain and the Netherlands. Science 313:351–354.

Breeze, T. D., A. P. Bailey, K. G. Balcombe, and S. G. Potts. 2011. Polli-nation services in the UK: how important are honeybees? Agriculture,Ecosystems and Environment 142:137–143.

Carvell, C., W. R. Meek, R. F. Pywell, and M. Nowakowski. 2004. Theresponse of foraging bumblebees to successional change in newly createdarable field margins. Biological Conservation 118:327–339.

Corbet, S. A. 1995. Insects, plants and succession: advantages of long-termset-aside. Agriculture, Ecosystems and Environment 53:201–217.

Corbet, S. A. 2000. Conserving compartments in pollination webs. Conserva-tion Biology 14:1229–1231.

Cranmer, L., D. McCollin, and J. Ollerton. 2011. Landscape structure influencespollinator movements and directly affects plant reproductive success(Early View). Oikos 121:562–568.

Dicks, L. V., S. A. Corbet, and R. F. Pywell. 2002. Compartmentalization inplant-insect flower visitor webs. Journal of Animal Ecology 71:32–43.

Dixon, K. W. 2009. Pollination and restoration. Science 325:571–573.Dormann, C. F., O. Schweiger, I. Augenstein, D. Bailey, R. Billeter, G. De

Blust, et al. 2007. Effects of landscape structure and land-use intensityon similarity of plant and animal communities. Global Ecology andBiogeography 16:774–787.

Dryden, R. 1997. Habitat Restoration Project: Fact Sheets and Bibliographies.English Nature Research Report No. 260, English Nature, Peterborough,United Kingdom.

Environment Agency for England and Wales. 2006. What are Landfill Sites?(available from http://www.environmentagency.gov.uk/yourenv/eff/1190084/resources–waste/213982/207743/?version=1&lang=e)[accessed on 8 November 2006].

Fitter, A. H., R. S. R. Fitter, I. T. B. Harris, and M. H. Williamson. 1995.Relationships between first flowering date and temperature in the floraof a locality in central England. Functional Ecology 9:55–60.

Forup, M. L., and J. Memmott. 2005. The restoration of plant-pollinatorinteractions in hay meadows. Restoration Ecology 13:265–274.

Forup, M. L., K. S. E. Henson, P. G. Craze, and J. Memmott. 2008. Therestoration of ecological interactions: plant-pollinator networks on ancientand restored heathlands. Journal of Applied Ecology 45:742–752.

Fussell, M., and S. A. Corbet. 1992. Flower usage by bumblebees: a basis forforage plant management. Journal of Applied Ecology 29:451–465.

Gallai, N., J. M. Salles, J. Settele, and B. E. Vaissiere. 2009. Economicvaluation of the vulnerability of world agriculture confronted withpollinator decline. Ecological Economics 68:810–821.

Gilbert, O. L., and P. Anderson. 1998. Habitat creation and repair. OxfordUniversity Press, Oxford, United Kingdom.

Goulson, D. 2003a. Bumblebees: their behaviour and ecology. Oxford Univer-sity Press, Oxford, United Kingdom.

Goulson, D. 2003b. Conserving wild bees for crop pollination. Food, Agricul-ture & Environment 1:142–144.

Handel, S. N., G. R. Robinson, and A. J. Beattie. 1994. Biodiversity resourcesfor restoration ecology. Restoration Ecology 2:230–241.

Hegland, S. J., and L. Boeke. 2006. Relationships between the density anddiversity of floral resources and flower visitor activity in a temperategrassland community. Ecological Entomology 31:532–538.

Holl, K. D. 1995. Nectar resources and their influence on butterfly communitieson reclaimed coal surface mines. Restoration Ecology 3:76–85.

Hutchings, T. R., D. Sinnett, A. J. Peace, and A. J. Moffat. 2006. The effectof woodland growth on a containment landfill site in Hertfordshire, UK.Urban Forestry & Urban Greening 5:169–176.

Ireland, E. M. 1991. Factors influencing the establishment of floristically richgrasslands on a restored landfill site. Ph.D. Thesis, Essex University,Colchester, United Kingdom.

Kalikhman, I. 2007. Patchy distribution fields: a spiral survey design andreconstruction adequacy. Environmental Monitoring and Assessment124:243–252.

Keith, S. A., A. C. Newton, M. D. Morecroft, C. E. Bealey, and J. M. Bullock.2009. Taxonomic homogenization of woodland plant communities over70 years. Proceedings of the Royal Society B: Biological Sciences276:3539–3544.

SEPTEMBER 2013 Restoration Ecology 567

Flowers and Pollinating Insects on Restored Landfills

Kent, D. H. 1992. List of vascular plants of the British isles. BSBI Publications,London, United Kingdom.

Kessel, S. R., and R. H. Whittaker. 1976. Comparisons of three ordinationtechniques. Vegetatio 32:21–29.

Klein, A. M., B. E. Vaissiere, J. H. Cane, I. Steffan-Dewenter, S. A.Cunningham, C. Kremen, and T. Tscharntke. 2007. Importance ofpollinators in changing landscapes for world crops. Proceedings of theRoyal Society B: Biological Sciences 274:303–313.

Kremen, C., N. M. Williams, R. L. Bugg, J. P. Fay, and R. W. Thorp. 2004.The area requirements of an ecosystem service: crop pollination by nativebee communities in California. Ecology Letters 7:1109–1119.

Legendre, P., and L. Legendre. 1998. Numerical ecology. Elsevier Science,Amsterdam, The Netherlands.

Maarel, E.v.d., editor. 2005. Vegetation ecology. Blackwell Publishing, Lon-don, United Kingdom.

McCollin, D., L. Moore, and T. H. Sparks. 2000. The flora of a culturallandscape: environmental determinants of change revealed using archivalsources. Biological Conservation 92:249–263.

McCune, B., and J. B. Grace. 2002. Analysis of ecological communities. MjMSoftware, Gleneden Beach, Oregon.

Meyer, B., F. Jauker, and I. Steffan-Dewenter. 2009. Contrasting resource-dependent responses of hoverfly richness and density to landscapestructure. Basic and Applied Ecology 10:178–186.

Natural England. 2009. What are Local Nature Reserves? (available fromhttp://www.naturalengland.org.uk/ourwork/conservation/designatedareas/lnr/default.aspx) [accessed on 23 February 2009].

Neal, P. R. 1998. Pollinator restoration. Trends in Ecology & Evolution13:132–133.

Nielsen, A., and J. Bascompte. 2007. Ecological networks, nestedness andsampling effort. Journal of Ecology 95:1134–1141.

Ockinger, E., and H. G. Smith. 2007. Semi-natural grasslands as populationsources for pollinating insects in agricultural landscapes. Journal ofApplied Ecology 44:50–59.

Ollerton, J., R. Alarcon, N. M. Waser, M. V. Price, S. Watts, L. Cranmer,A. Hingston, C. I. Peter, and J. Rotenberry. 2009. A global test of thepollination syndrome hypothesis. Annals of Botany 103:1471–1480.

Ollerton, J., R. Winfree, and S. Tarrant. 2011. How many flowering plants arepollinated by animals? Oikos 120:321–326.

Osborne, J. L., and I. H. Williams. 1996. Bumble bees as pollinators of cropsand wild flowers. Pages 24–32 in A. Matheson, editor. Bumble beesfor pleasure and profit. International Bee Research Association, Cardiff,United Kingdom.

Osborne, J., A. Martin, N. Carreck, J. Swain, M. Knight, D. Goulson, R. Hale,and R. Sanderson. 2008. Bumblebee flight distances in relation to theforage landscape. The Journal of Animal Ecology 77:406–415.

Pollard, E., and T. J. Yates. 1993. Monitoring butterflies for ecology andconservation. Chapman & Hall, London, United Kingdom.

Potts, S. G., T. Petanidou, S. Roberts, C. O’Toole, A. Hulbert, and P. Willmer.2006. Plant-pollinator biodiversity and pollination services in a complexMediterranean landscape. Biological Conservation 129:519–529.

Potts, S. G., J. C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, and W.E. Kunin. 2010. Global pollinator declines: trends, impacts and drivers.Trends in Ecology & Evolution 25:345–353.

Ricketts, T. H., J. Regetz, I. Steffan-Dewenter, S. A. Cunningham, C. Kremen,A. Bogdanski, et al. 2008. Landscape effects on crop pollination services:are there general patterns? Ecology Letters 11:499–515.

Stace, C. 1991. New flora of the British Isles. Cambridge University Press,Cambridge, United Kingdom.

Steffan-Dewenter, I., and T. Tscharntke. 2001. Succession of bee communitieson fallows. Ecography 24:83–93.

Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tscharntke.2002. Scale-dependent effects of landscape context on three pollinatorguilds. Ecology 83:1421–1432.

UK National Ecosystem Assessment. 2011. Synthesis of the Key Find-ings. UNEPWCMC, Cambridge, United Kingdom (available fromhttp://uknea.unep-wcmc.org/) [accessed 2 November 2011].

Watson, D., and V. Hack. 2000. Wildlife management and habitat creation onlandfill sites: a manual of best practice. Ecoscope Applied Ecologists,London, United Kingdom.

Winfree, R., N. M. Williams, J. Dushoff, and C. Kremen. 2007. Native beesprovide insurance against ongoing honey bee losses. Ecology Letters 10:1105–1113.

Wong, M. H. 1988. Soil and plant characteristics of landfill sites nearMerseyside, England. Environmental Management 12:491–499.

568 Restoration Ecology SEPTEMBER 2013