CONSERVINGFARMLANDBIODIVERSITY

TEAGASCBiodiversityConference

Lessons learned & futureprospects

25-26 MAY 2011

PROCEEDINGS

Published by

Teagasc,Head Office,Oak Park,Carlow

www.teagasc.ie

May 2011

Teagasc Biodiversity Conference

ConservingFarmland

Biodiversity

Lessons learned & future prospects

Edited by Daire Ó hUallacháin and John Finn

A Chara,

On behalf of the Conference Committee, I would like to extend a warm welcome to everyoneattending this Teagasc conference ‘Conserving Farmland Biodiversity: lessons learned andfuture prospects’.

For the European Union and its Member States, biodiversity continues to be a keyenvironmental objective. The EU target to halt biodiversity loss by 2010 has not been met,and the EU is strengthening its policy framework and commitment to halting the loss ofbiodiversity and the degradation of ecosystem services in the EU by 2020, and restoring themin so far as possible. Thus, the success of biodiversity measures (in a variety of policies andsectors) will increasingly be judged by the extent to which they halt the loss of biodiversityand halt the degradation of related ecosystem services. In the context of the agriculture sector,these policy commitments are likely to be strengthened in the post-2013 CAP, although thereremains considerable uncertainty about the specific policy instruments, measures andespecially funding that will be used to achieve them.

Naturally, these European-scale policy commitments will be translated into nationalcommitments. In its recent Government strategy for agriculture, the Food Harvest 2020 reportoutlines a vision for the agri-food sector in which the conservation of biodiversity appears asone of the priority environmental goals. The updated Irish Biodiversity Action Plan 2010-2015 also comes into effect this year and, in addition to the protection of designated areas(and other targets), clearly highlights the importance of biodiversity conservation in the widercountryside, most of which is farmland.

The problem with the wider countryside is that it is so wide! In addition, it is highly variedwith biodiversity being distributed in a very uneven way. Some farmland areas contain a levelof biodiversity that rivals the quality of that in Natura 2000 sites and other designated areas;others support a lower (but still valued) level of biodiversity that persists within pockets ofsemi-natural habitats that interact with and are located within a wider matrix of moreintensively-managed farmland. This variation generates a number of questions: How effectivehave previous conservation initiatives been in conserving farmland biodiversity, and what arethe drivers of success or failure? What relative emphasis will policies place on theconservation of designated sites and the wider countryside, and the different objectives ofhabitat protection, restoration and creation? Do we have sufficient information on thedistribution of biodiversity (and its threats) across the wider countryside and, if not, how bestto get it? How can farmers appropriately manage specific habitats and species whilstproducing food and making a livelihood? What is the ‘best’ allocation of limited budgets forconservation between higher- and lower-quality habitats in the wider countryside? Thesequestions are deceptively simple, but the answers are certainly not.

We hope that this conference contributes to progressing efforts toward addressing thesequestions, and look forward to the conference being as enjoyable as it will be informative.

On behalf of the Conference Committee:Dr John Finn (Chair)Dr Daire Ó hUallacháinMr Pat MurphyMs Catherine KeenaMr Stuart Green

Wednesday 25th May09.00-10.00 Conference registration

10.00-11.10 Session 1

10.00 Dr Noel Culleton Introduction and welcome

10.10 Keynote: Dr David Baldock

Environmental public goods and the post-2013 CAP

10.50 A. Bleasdale Agri-environmental policy perspectives ofthe National Parks and Wildlife Service

11.10-11.40 Coffee (poster session)

11.40-13.20 Session 2

11.40 Keynote: Prof Nick Sotherton

Arable management options in the UK agri-environmentscheme: the research behind the options

12.20 M. Jebb Important areas of plant diversity outside ofdesignated areas: where are they?

12.40 D. Doody A critical source area approach to thedevelopment of supplementary measures forhigh-status waterbodies

13.00 Panel discussion

13.20-14.20 Lunch

14.20-15.30 Session 3

14.20 Keynote: Dr Sharon Parr

The Burren Farming for Conservation Programme: lessonslearned and progress to date

15.00 Three minute poster presentations

15.30-15.50 Coffee (poster session)

15.50-17.30 Session 4

15.50 Keynote: Dr Guy Beaufoy

HNV farming policy – progress to-date & future challenges

16.30 J. Finn The environmental impact of REPS: lessonslearned & future prospects

16.50 N. Smyth Invasive species in Ireland

17.10 Prof. Gerry Boyle Summary and panel discussion

20.00 Conference dinner

Thursday 26th May09.00-10.40 Session 5

09.00 Keynote: Dr Simon Mortimer

Biodiversity options in agri-environment schemes for moreintensive farmers

09.40 R. Fritch Improving floral and invertebrate diversity ingrassland field margins

10.00 C. Sullivan What is the conservation potential ofgrasslands on lowland farms?

10.20 J. Martin High nature value meadows: results from anational grassland survey

10.40-11.00 Coffee (poster session)

11.00-13.00 Session 611.00 Keynote: Mr Daniel Fuchs

Identification and distribution of HNV farmland inGermany

11.40 H. Sheridan Habitats in the Irish farmed landscape

12.00 J. Stout Pollinators and pollination networks in Irishfarmland: implications for conservation ofpollination services

12.20 J. McAdam Monitoring previous agri-environmentschemes in Northern Ireland- a review

12.40 Panel discussion

13.00-14.00 Lunch

14.00-15.10 Session 714.00 Keynote: Dr Evelyn Moorkens

Evidence-based selection of catchments and measures forconservation and recovery of freshwater pearl musselpopulation in Ireland

14.40 Three minute poster presentations

15.10-15.30 Coffee (poster session)

15.30-16.30 Session 815.30 K. Buckley The role of habitat creation in the recovery

of the Irish grey partridge Perdix perdix

15.50 A. Copland REPS and farmland bird populations: resultsand recommendations from the FarmlandBirds Project

16.10 Panel discussion

16.30 Close of conference

Contents

ORAL PRESENTATIONSAgri-environmental policy perspectives of the National Parks 1and Wildlife ServiceA. Bleasdale and M. Dromey

Arable management options in the UK Agri-environment scheme: 3the research behind the optionsN.W. Sotherton

Important areas of plant diversity outside of designated areas: 6where are they?J.A. Finn and M. Jebb

A critical source area approach to the development of supplementary 8measures for high status waterbodiesD.G. Doody & R.P.O Schulte

HNV farming policy – progress to-date and future challenges 10G. Beaufoy

Evidence on the environmental effectiveness of REPS: 15a literature reviewJ.A. Finn and D. Ó hUallacháin

Invasive species control case study: 17Hottentot fig (Carpobrotus edulis) control in Ireland.N. Smyth, M. Jebb, and A. Booth

Options for enhancing the biodiversity value of intensive livestock farms: 19experience from the English agri-environment schemesS.R. Mortimer

Improving floral and invertebrate diversity in grassland field margins 24R. Fritch, H. Sheridan, J.A. Finn, and D. ÓhUallacháin

What is the conservation potential of grasslands on lowland farms? 26C.A. Sullivan, M. Sheehy Skeffington, J.A. Finn and M.J. Gormally

High nature value meadows: results from a national grassland survey 28J.R Martin and F.H. O’Neill

Identification and distribution of HNV farmland in Germany 30D. Fuchs and A. Benzler

Habitats in the Irish farmed landscape 32H. Sheridan, B. Keogh, A. Anderson, T. Carnus, B.J. McMahon,S. Green and G. Purvis

Pollinators and pollination networks in Irish farmland: 34implications for conservation of pollination servicesJ.C. Stout, E.F. Power, D.A. Stanley and S.E. Mullen

Monitoring of the Agri-environment Programme within the 36Northern Ireland Rural Development Programme 2007-2013M. Flexen, D. O’Mahony and J.H. McAdam

Linking catchment activities with conservation status of freshwater 38pearl mussel populations as an aid to prioritization of rehabilitationmeasures in IrelandE.A. Moorkens and S. Downes

The role of habitat creation in the recovery of the Irish grey partridge 42Perdix perdixK. Buckley, C. O’Gorman, P. Comerford, V. Swan, C. Flynn, T. Carnus,B. Kavanagh and B.J. McMahon

REPS and farmland bird populations: results and recommendations from the 44Farmland Birds ProjectA.S. Copland and J. O’Halloran

POSTERSLearning to improve agri-environmental schemes: using experts’ judgements 46to assess the environmental effectiveness of the Rural EnvironmentalProtection SchemeD. Bourke, I. Kurz, L. Dunne and J.A. Finn

Halting biodiversity loss: the potential of High Nature Value farming in 48north-west IrelandP.G. Boyle, M. Hayes-Fitzpatrick, M. Gormally and J. Moran

Upland commonages in Connemara – can biodiversity be used to inform 50sustainable agri-environmental policy?B. Canning, A. Bell, B. Mongan, M. Gormally, M. Sheehy Skeffington andT. van Rensburg

Probability of pollinator occurrence increases with duration of AES participation 52and floral diversity

T. Carnus, A. Anderson, J. Breen, B.J. McMahon, V. Santorum, H. Sheridanand G. Purvis

Positive effects of sown evenness on sward biomass across three years, 54nitrogen application and cutting intensityT. Carnus, J.A., Finn, A. Cuddihy, L. Kirwan and J. Connolly

Conserving Ireland’s farmland birds: performance and prospects of 56agri-environment schemesA.S. Copland, O. Crowe, J. O’Halloran and A.W. Lauder

The food versus fuel debate – what effect will replacing traditional crops 58with Miscanthus x giganteus have on farmland biodiversity?M. Emmerson, D. Bourke, J. Dauber, E. O’Rourke, D. Stanley, R. Thompson,P. Whelan and J. Stout

How to measure the environmental effectiveness of the Rural Environmental 60Protection Scheme?J.A. Finn

Landscape aesthetics: assessing the general publics’ preferences for 62rural landscapesP. Howley

Cessation of grazing - the effects on land snail communities in farmed grassland, 64scrub and woodland habitats in the Burren region in the west of IrelandM.P. Long, E.A. Moorkens and D.L. Kelly

Development and application of the Agri-environmental Footprint Index (AFI) 66G. Louwagie, J.A. Finn, G. Purvis and G. Northey

Can current Irish riparian management practices enhance ground beetle diversity? 68D. Madden, S. Harrison, J.A. Finn and D. Ó hUallacháin

Do cattle drinking points contribute to stream pollution levels in Ireland? 70D. Madden, D. Ó hUallacháin, J.A. Finn and S. Harrison

Hay meadow plant communities on the Shannon Callows: responses to summer 72flooding and changes in managementC.A. Maher, M.J. Sheehy Skeffington and M.J. Gormally

Development and design of locally targeted HNV programmes in Ireland – 74The Aran Islands Case StudyP. McGurn and J. Moran

Conserving Twite Carduelis flavirostris in Ireland 76D.T. McLoughlin and D.W. Norriss

How does animal stocking rate influence agricultural grassland biodiversity 78when measured at different scales?B.J. McMahon, A. Anderson, T. Carnus, A.J. Helden, H. Sheridan and G. Purvis

Lethal and sub-lethal effects of ivermectin on two common species of dung 80beetle in a laboratory experimentN.M. O’Hea, L. Kirwan, P.S. Giller and J.A. Finn

Experts’ assessments of biodiversity options and supplementary measures 82in REPS 4D. Ó hUallacháin, J.A. Finn, M. Gormally, and C. Carlin

HNV farming and the management of upland biodiversity on the 84Iveragh Peninsula, South KerryE. O’Rourke

Mapping the broad habitats of the Burren using satellite imagery 86S. Parr, J.A. Finn and G. O’Donovan

The National Survey of Upland Habitats 88P.M. Perrin, J.R. Roche, O.H. Daly and C. Douglas

Floral resources for bumblebees on Irish farmland 90V. Santorum and J. Breen

Agri-environment management of grassland field margins 92S.E. Spratt, A. Cooper and T. McCann

Mapping High Nature Value farmland in Ireland 94C.A. Sullivan, D.A. Bourke, J.A. Finn, M. Sheehy Skeffington and M. J. Gormally

Impact of riparian vegetation structure on small mammal communities 96D. Ó hUallacháin and D. Madden

OralPresentations

Agri-environmental policy perspectives ofthe National Parks and Wildlife Service

A. Bleasdale and M. DromeyNPWS, Dept of Environment, Heritage andLocal Government, 7 Ely Place, Dublin 2E-mail: andy.bleasdale@environ.ie

IntroductionIreland has significant obligations under theHabitats and Birds Directives which require us to designate SACs and SPAs (the Natura

2000 network), to maintain, and in many cases improve,

the status and condition of habitats andspecies in designated areas (which willrequire measures inside, and sometimesoutside, these sites),

to protect national populations of naturallyoccurring bird species (in the widercountryside).

As the designation process is substantiallycompleted, the focus is now on the protectionand management of sites, habitats and species.As most of the lands of the State are managedthrough farming, agricultural policy, schemesand initiatives are central to the conservationof farmland biodiversity.

Financial supports for biodiversity-focusedfarming are critical, including, but notconfined to, agri-environmental schemes underPillar II of the of the Common AgriculturalPolicy (CAP). In addition, direct (Pillar I)supports to farmers are conditional onadherence to the Birds and Habitats Directives,under cross compliance. Appropriate targetingof future funding, provided through thenational Rural Development Programme 2014-20 (RDP), is critical if farmland biodiversity isto be protected in designated areas and in thewider countryside.

Policy DriversIreland’s performance in relation to theappropriate management of Natura 2000 siteshas been assessed by the EU through a numberof European Court of Justice infringementcases. To date, it is clear that agriculturalsupports have not always ensured adequateprotection of biodiversity. Ireland must reportto the EU in 2013 under Article 17 of theHabitats Directive on the “conservation status”

for a suite of habitats (and species supportedby these habitats). This report is likely toindicate that farming has a significant role toplay in delivering “favourable conservationstatus”. The challenge is to ensure thatagriculture and conservation policies aresufficiently aligned to achieve this goal.

In the coming years, Ireland will be expectedto address the following issues: management of freshwater systems, conservation of the freshwater pearl

mussel, management of overgrazed (and

undergrazed) uplands, conservation of birds in serious decline,

including species associated withagriculture such as the corncrake, breedingwaders, partridge, barn owl etc.,

protection and restoration of bogs andother wetlands.

As many of the above issues are influenced byfarming systems and supports, it is imperativethat all policies are “Natura 2000 proofed”, toensure appropriate management of thisbiodiversity resource. This will require thetailoring of direct supports and agri-environment schemes (and measures) toprovide the best solutions to meet Ireland’sobligations.

Research GapsThe Irish National Platform for BiodiversityResearch (NPBR) is a forum for policymakers, scientists and other interestedstakeholders involved in the field ofbiodiversity research. The platform is rununder the auspices of NPWS and theEnvironmental Protection Agency (EPA) andis administered by a secretariat. The aim of theplatform is to facilitate the targeting and co-ordination of biodiversity research in Irelandto meet the current and future policy needs. Tothis end, a cross-sectoral Agriculture WorkingGroup was established to identify key researchquestions. In synopsis, the emerging researchgaps that need to be urgently addressed are asfollows (NPBR, 2011): national inventories of species and habitats

supported by agricultural systems,including remedial actions to protect them,

an investigation of the effect of variousmanagement techniques in a range offarmed habitats,

an investigation of the impacts ofalternative land-uses in marginalagricultural areas,

guidance on the future grazing ofextensive upland habitats (includingcommonages),

a description of the agricultural systemsthat support High Nature Value (HNV)farmland,

an assessment of the socio-economics offarming for conservation,

a description of the ecosystem services ofdifferent farming systems (with differentassociated biodiversity attributes),

development of tailored, testedprescriptions for roll-out through AEschemes,

development of indicators for monitoringat farm and scheme level,

the establishment of long-term study sites.

DiscussionArticle 8 of the Habitats Directive provides forfunding arrangements towards the cost ofmanaging priority species and habitats. TheEU Commission takes the view that funds aremade available through, and should be sourcedfrom, the existing instruments, in particular theCAP supports.

A substantial part of funding under the currentRDP is earmarked for spending related to theenvironment, in particular in the areas ofbiodiversity, water protection and climatechange mitigation. From an NPWSperspective, the implementation of Pillar I andPillar II payments to farmers have not, to date,always ensured that farmland biodiversity hasbeen adequately protected.

An appropriate allocation of CAP funding inthe next cycle (2014-2020) is critical todelivery of some of Ireland’s obligations underthe Habitats and Birds Directives. NPWSsupports the continuation of agri-environmentsupports from Pillar II, but greater targeting ofthese supports, to meet some of the challengesidentified previously, is required. This willnecessitate cross-departmental co-operationbetween policy makers.

The HNV farmland concept is one that shouldbe advanced through the CAP and the nationalRDP. Low-intensity, extensive farming allowsbiodiversity to flourish and can result in HNVtype farmland outside of designated areas

(Smith et al., 2011). Through the proposed“greening” of Pillar I, NPWS would like to seedirect financial supports in place that wouldincentivise the production of “biodiversityadded value” (i.e. HNV farmland).

The Burren Farming for ConservationProgramme (BFCP) model shows that, with atailored and tested prescription and with theclose co-operation of departments and thefarming community, real advances can bemade in the area of biodiversity protection andenhancement (Dunford, 2011). This model isadmired throughout Europe, has been shownto work, meets all the requirements of HNVfarmland and will contribute towardsfavourable conservation status within SACs.There are lessons to be learned here that couldbe usefully applied in other parts of thecountry.

ConclusionIt is imperative that payments and supports tofarmers are targeted to address issues of EUand national concern, including the protectionof Natura 2000 sites and wider biodiversityand ecosystem services.

AcknowledgementsWe are grateful to Louise Scally (BECConsultants Ltd, Secretariat of the NPBR) andthe members of the NPBR AgricultureWorking Group, which includedrepresentatives from the EPA, NPWS, DAFF,the Heritage Council, Teagasc, UCD and theSligo Institute of Technology.

ReferencesDunford, B. (2011). BFCP Year 1 Report.http://www.burrenlife.com

NPBR (2011). Research recommendations ofthe Agriculture, Grasslands and Soil WorkingGroup of the National Platform forBiodiversity Research.

http://www.biodiversityresearch.ie

Smith, G.G., O’Donoghue, P. and O’Brien, C.(2010). Sustaining High Nature Farmingsystems; lessons from the west of Ireland. InPractice 71: 22-25.

Arable management options in the UK Agri-environment scheme: the research behind theoptions

N.W. SothertonGame & Wildlife Conservation Trust,Fordingbridge, Hampshire, SP6 1EFEmail: nsotherton@gwct.org.uk

IntroductionThe Game & Wildlife Conservation Trust’s(GWCT) involvement in agroecology dates back tothe 1930s when concerns for the demise of thepopulation status of the UK’s most abundantgamebird, the grey partridge (Perdix perdix),became apparent. Over the last 40 years, thisspecies has declined by 88% (Marchant et al.,1990) and reduced its range by 19% (Gibbons etal., 1993). Of all the farmland birds, greypartridges are the species most closely associatedwith the farmed landscape and have thereforebecome the iconic farmland bird. Because of ourinterest in grey partridges, we were acutely awareof the plight of farmland wildlife and the problemscaused by intensive agriculture decades before therest of the UK conservation community woke up tothis problem and before the Biodiversity ActionPlan process was initiated in the mid 1990s. To thisday, the grey partridge remains the best studiedfarmland bird.

By this time we had already conducted the large-scale, multi-year experiments to unravel the causesof decline and indicate the mitigation measuresthat should be taken to initiate population recovery(Sotherton, 1991; Tapper et al., 1996). Themajority of these mitigations involved habitatcreation and management schemes to createessential nesting cover, insect-rich brood cover tofeed young chicks and winter cover to providefood and shelter from weather and protectedpredators, especially raptors.

When agri-environment schemes (AES) werebeing developed, the GWCT already had a menuof management prescriptions available forincorporation into a UK AES and for immediateuse. Of the 36 options available in the AES inEngland six were entirely developed by the GWCTand in a further 23, our research had played amajor role in their scientific evaluation.

Arable PrescriptionsSuch prescriptions developed by the GWCT,where the rigorous scientific evidence of efficacyis in place, are listed below. We believe that suchan evidence base to support managementprescriptions is an absolute prerequisite for an AESto give confidence to users and funders alike.

Conservation headlandsThe use of selective pesticides, both herbicides andinsecticides, along the edges of arable crops toprotect the flora and the invertebrate life per se butthat includes the chick-food insect groups that feedyoung birds in the first few weeks of life (Potts,1986). Selectivity allows pernicious weeds, nottolerated by farmers, to be removed whilst leavingkey insect host plants. This is a compromise thatworks.

Conservation headlands remaining unharvestedThese are designed to feed young chicks during thesummer. The unharvested, selectively sprayedcereal crop edges are left uncut to provide food andshelter into the winter.

Conservation headlands with low rates of fertiliserThese are designed to protect our rare arable floraand associated invertebrates.

Grassy marginsThese are from two to six metres wide and providenesting cover for ground-nesting birds. They alsosupport many groups of wildlife and bufferboundary habitats and watercourses from adjacentagricultural operations.

Beetle banksRaised grassy banks of perennial grasses sownacross the centres of large arable fields to breakthem up, providing habitats for a more diverseinvertebrate fauna and nest sites for ground-nesting birds and harvest mice.

Wild bird cover mixturesAdapted from the concept of game cover crops,these are sown blocks of seed-bearing cropsdesigned to feed and shelter birds in winter andespecially see birds through the hungry gap.

Secrets of SuccessAs a Non-Government Organisation and registeredcharity, how have we been so influential in helpingdetermine Agri-environment policy?

Evidence basedEvidence for efficacy of our management optionswas in place and published in peer-reviewedjournals.

CostedWe could calculate levels of compensation neededor amounts of profits foregone.

Farmer-friendlyOur scientists have farming backgrounds so ourprescriptions were workable, simple and

achievable by a farming workforce. There must beflexibility over pesticide and fertiliser use whenmanaging land in an AES prescription. If growinga crop of kale (to produce seed to feed farmlandbirds in winter), applying 80kg/ha of nitrogenfertiliser produces far more seed than whenapplying 40kg or less. Also farmers are concernedwhen Agri-environment habitats revert to plots ofpernicious weeds. Selective herbicides on a pre-approved list should be available to target specificweeds and give farmers confidence to farm forwildlife.

On-going researchManagement options are constantly reviewed andrenewed, especially with regard to problemsidentified by farmers. Regular feedback is given tothe authorities to fine-tune or tweak the scheme.

Clear objectivesSpecific targets of species or groups are identifiedand management specifically targeted at them.Such targets in arable ecosystems include greypartridge, farmland birds (species that also feedtheir chicks on insects), chick-food insects, annualarable wildflowers. Rarity, BAP status, knowledgeof successful mitigation and ecosystem functionhave informed this choice.

Grassland PrescriptionsArguably, wildlife losses in grassland farminglandscapes in the UK have been as great if notgreater than losses from arable land. Our limitedexperience of designing agri-environmentprescriptions for grass farms, especially intensivedairy enterprises, is that it is difficult andexpensive. All schemes need stock protection andtherefore will require expensive fencing. Specificgrassland measures will be described elsewhere byother key speakers but wildlife will benefit fromsome of the arable prescriptions transposed into agrassland landscape, if only to recreate the mixedarable/grassland farming system that would haveonce been more common.

The FutureMany questions have been asked about the efficacyof the agri-environment measures across Europe(Kleijn & Sutherland, 2003). We believe they areworking, but they could be delivering moreenvironmental protection and be more effective atspecies recovery. To do this, the next generation ofschemes in the post-2013 CAP reform should:(a) Be better targeted to species relevant to that

area so, for example, measures do not supportapplications to support rare arable chalk florason clay soils.

(b) Improve the payment rates for theseprescriptions that are more difficult to

implement to discourage farmers opting for theeasy options. For example, in England, thetarget for permanent grass margins aroundarable fields has been exceeded by nearlythree-fold, whereas the more taxing targets forflower-rich margins, low-input field marginsand wild bird seed mixtures have all failed toreach their 2010 targets. Permanent grassmargins are easy, qualify for relativelyattractive rates of payment and are easy tomaintain. Therefore they are over-subscribed.Also, schemes should constantly reviewpayment rates to account for the volatility ofmarkets and the popularity and uptake of themost valuable prescriptions.

(c) Make the prescriptions multifunctional. So farwe have achieved prescriptions for six-metrewide strips of nesting cover, brood cover andwinter cover. With wheat selling at€200/tonne, uptake of Agri-environment maybe low and so what Agri-environment land wehave will have to deliver more. Designingschemes for the management of such land willbe the job of research scientists.

(d) Predator control should be an integral part ofAgri-environment schemes and paymentsshould be made available for those prepared todo it. We do not doubt that species declineswere caused by a changing and increasinglyintensive agriculture. Declines were not causedby predation. However, we increasinglybelieve that species recovery in the presence offirst class habitat management may be madedifficult or impossible by predation. In the UKthere have been big increases in common,ubiquitous generalist predators such as foxes,crows and rats. Experimental removal of thesepredators during the nesting season to protectthe sitting hen and their nests leads to bigincreases in productivity and subsequentbreeding densities (Tapper et al., 1996;Fletcher et al., 2010). The same is true for thebrown hare (Lepus europaeus) where foxcontrol protects the leveret productionextremely well (Reynolds et al., 2010).

(e) Management prescriptions should come inbundles. If your target is species recovery, thenall the elements (not just some) of the recoverypackage must be chosen. “Cherry picking”some options but not others from a packageshould not be permitted. For example, if youchose to recover your grey partridgepopulation you would need to opt for theprovision of all the covers needed (nesting,brood, winter) and predator control.

(f) AES deliver more when the farmer hasreceived specialist advice compared to wherehe/she has not.

(g) Given the scale at which farmland wildlifeoperates, greater consideration and supportshould be given to applications where farmershave worked together within a district orparish.

(h) The success of a scheme should be judged byoutputs, not numbers of participants/schemessigned up. 70% of UK eligible land is in ascheme, but many species of farmland still failto recover. Perhaps we start with the wrongquestion? Farmers are asked to join a schemenot “what wildlife would you like to see backon the farm”?

ConclusionsElsewhere in this conference (see contributionfrom Buckley et al.), you will be able to learn moreabout the National Grey Partridge ConservationProject in County Offaly. There, using all thepotential elements of an Agri-environment scheme,the NPWS and others have saved the most westerlypopulation of grey partridges in Europe fromextinction. Their success should be a model forothers. In the UK, I believe Agri-environment isworking but I am even more certain that we can dobetter.

AcknowledgmentsWe thank Teagasc for their invitation to attend thisconference.

ReferencesFletcher, K., Aebischer, N.J., Baines, D., Foster, R.& Hoodless, A.N. (2010). Changes in breedingsuccess and abundance of ground-nestingmoorland birds in relation to the experimentaldeployment of legal predator control. Journal ofApplied Ecology, 47: 263-272.

Gibbons, D.W., Reid, J.B. & Chapman, R.A.(1993). The New Atlas of Breeding Birds inBritain and Ireland: 1988-1991. T. & A.D. Poyser,London.

Kleijn, D. & Sutherland, W.J. (2003). Howeffective are European agri-environment schemesin conserving and promoting biodiversity? Journalof Applied Ecology, 40: 947-969.

Marchant, J.H., Hudson, R., Carter, S.P. &Whittington, P. (1990). Population trends in BritishBreeding Birds, British Trust for Ornithology,Tring.

Potts, G.R. (1986). The Partridge: Pesticides,Predation and Conservation, London, Collins.

Reynolds, J.C., Stoate, C., Brockless, M.H.,Aebischer, N.J. & Tapper, S.C. (2010). The

consequences of predator control for brown hares(Lepus europaeus) on UK farmland. EuropeanJournal of Wildlife Research, 56: 541-549.

Sotherton, N.W. (1991). Conservation Headlands:a practical combination of intensive cereal farmingand conservation. In Firbank, L.G., Carter, N.,Derbyshire, J.F.

Tapper, S.C., Potts, G.R. & Brockless, M.H.(1996). The effect of an experimental reduction inpredation pressure on the breeding success andpopulation density of grey partridges Perdixperdix. Journal of Applied Ecology, 33: 965-978.

Important areas of plant diversity outside ofdesignated areas: where are they?

J.A. Finn1 and M. Jebb2

1Teagasc, Environment Research Centre,Johnstown Castle, Wexford2National Botanic Gardens, Glasnevin, Dublin,Email: john.finn@teagasc.ie

IntroductionHaving failed to halt biodiversity loss by 2010, theEU has strengthened its commitment to preventfurther biodiversity loss by 2020. Biodiversitygoals will not be met solely from protection ofdesignated nature conservation sites, and theconservation of biodiversity on farmland outside ofprotected areas (e.g. Natura 2000 sites) will becritical to halting the loss of biodiversity (Jacksonet al., 2009). Halting the loss of biodiversity willrequire targeting of effort towards those speciesmost at risk of extinction or most threatened.Achieving this policy objective will requireinformation on the spatial distribution of importantareas for biodiversity that occur outside ofprotected areas.

Here, we describe a project that aims to resolvethis knowledge gap (whilst acknowledging thatthere are several other projects aiming to do soalso). This research is timely (and now time-critical) to provide the evidence base to anticipateand address more stringent policy requirements forenvironmental protection, and to facilitate thespatial targeting of agri-environment payments thatwill be required in the post-2013 CommonAgricultural Policy, and by Ireland’s NationalBiodiversity Plan 2010-2015.

Materials and MethodsWe will describe and prioritise the biodiversityvalue of different areas within the 26 counties ofthe Republic of Ireland, based on objective andtransparent criteria that reflect a hierarchy ofimportance of plant diversity. In this way, we willidentify important areas of plant diversity withineach county of the Irish countryside. We willcompare the distribution of these hotspots with thedistribution of protected areas (i.e. Natura 2000sites).This project will collate available data on thedistribution of individual rare plant species(available from National Botanic Gardens,Ireland). This represents over 900,000 records forplants of conservation interest at the scale ofhectads (10km x 10km squares) and about 135,000records at the scale of tetrads (2km x 2km squares).Here, we examined the hectad-scale distribution ofeach of the 63 named Flora Protection Orderspecies (FPO, 1999), and examined how many of

the hectads coincided with an area of Special Areaof Conservation (SAC).

Using available records of rare plants in Co.Leitrim (at a 2km x 2km resolution), we collatedplant species occurrence data for the followinggroups (at least); legally protected species (FPO,1999; CEC, 1992), nationally threatened species(Curtis and McGough, 1998], axiophytes* andindicator plant species that identify Annex Ipriority habitats (CEC, 1992). A provisionalcounty-level map of plant biodiversity hotspotswas plotted. We plotted the tetrads according to anumber of rules, in approximate order ofbiodiversity importance: contain at least one FPOspecies; contain at least one Red Data book plantspecies; contain 30 axiophyte species; contain atleast 50% of the indicator species for priority EUhabitats (not listed here). These methods are basedupon preliminary work undertaken for the Flora ofCounty Waterford (Fitzpatrick and Green, in prep).[*’Axiophytes’ refer to plant species of particularbotanical interest, they are usually indicators ofstable, diverse sites of botanical conservationvalue.]

Results and DiscussionThe number of hectads occupied by the 63 FPOspecies ranged from 1 hectad (five species: Carexdepauperata Starved Wood Sedge, Callitrichetruncata Short-leaved Water-Starwort, Hydrillaverticillata Irish Hydrilla, Minuartia recurvaRecurved Sandwort, Saxifraga nivalis Alpinesaxifrage) to 113 hectads (Omalotheca sylvatica,wood cudweed), and had a median of 10 hectads.Across each species, there was very large variationin the percentage of hectads with an FPO speciesthat also contained an area of SAC. Based on allrecords, 32 of the 63 species had 25% of theirdistribution outside of a hectad that coincided withan area of SAC. For 14 species, 50% of thehectads in which they occurred were not located ina hectad that coincided with an area of SAC. Notethat a designated site (SAC) only needed to occurin part of a hectad for the whole hectad to beidentified as a ‘designated’ hectad. Thus, thismethod certainly overestimates the extent to whichthe distribution of Flora Protection Order speciescoincide with SAC areas. The role of agri-environment schemes in protecting these speciesoutside protected areas may be of criticalimportance to their survival.

In a provisional analysis of records from Co.Leitrim, a total of 43 tetrads (out of a county totalof about 400 tetrads) were identified as importantareas of plant diversity (note that some tetradssatisfied more than one criterion). Twenty-twotetrads contained at least one FPO species; twelve

contained at least one Red Data book plant species;five contained 30 axiophyte species, and;fourteen contained at least 50% of the indicatorspecies for priority EU habitats.

At present the number of plant records (900,000)in the data set represents scarcely 13 records perkm2, and some areas of the country have beenmore intensively surveyed than others. Wetherefore analysed a number of different countiesfrom which the plant record density varied by twoorders of magnitude (Waterford, Leitrim andGalway). We found that the methods gave acomparable number of tetrads for each. Thissuggests that even when sparse, the records for rareor interesting species tend to be most thorough,whilst species of least concern are omitted. It islikely that gathering more data will give adiminishing return on such an investment(Grantham et al., 2008) and that the present dataset will prove robust.

Fig. 1. Preliminary application of methodology todistribution of records of rare plant species in Co.Leitrim. Each point represents multiple records ofdifferent types of rare plant species. In order ofdecreasing importance: Flora Protection OrderSpecies (red); Red Data Book species (blue);species of conservation value (green); indicatorspecies for Priority habitats (yellow). Note thatsome tetrads satisfied more than one criterion.County Leitrim has about 12,500 tetrad-levelrecords. Further work could, for example, comparethese distributions with the distribution of Natura2000 sites and participation in agri-environmentschemes.

ConclusionsThese data on Flora Protection Order speciesindicate that a considerable distribution of thesespecies occurs outside of designated areas. Haltingbiodiversity loss of these most threatened speciescould be facilitated by agri-environment measuresthat are highly spatially targeted in the widercountryside. The approach proposed here could beused to identify and protect important areas forplant diversity, and could contribute to effectivespatial targeting of agri-environment schemes.

AcknowledgmentsThis work draws upon the considerable efforts ofmany skilled voluntary recorders of plant recordsover more than a century, and the BotanicalSociety of the British Isles is thanked for providingthe data.

ReferencesCEC (1992) Council Directive 92/43/EEC on theconservation of natural habitats and of wild faunaand flora. The Council of the EuropeanCommunities: Luxembourg.

Curtis, T.G.F. and McGough, H.N. (1988) TheIrish red data book. Dublin: Wildlife Service,Ireland.

FPO. (1999) Flora Protection Order, in S.I. No. 94of 1999.

Fitzpatrick, Ú. & Green, P., Axiophytes andimportant habitats, www.nbdc.ie. (in press)

Grantham, H.S. et al. (2008) Diminishing return oninvestment for biodiversity data. ConservationLetters, 1: 190–198.

Jackson, S.F., Walker, K. and Gaston. K.J. (2009)Relationship between distributions of threatenedplants and protected areas in Britain. BiologicalConservation, 142: 1515-1522.

A critical source area approach to thedevelopment of supplementary measures forhigh status waterbodies

D.G. Doody 1 & R.P.O Schulte2

1Agri-Food and Bioscience Institute, Belfast,Northern Ireland, 2Teagasc Johnstown CastleWexfordEmail: donnacha.doody@afbini.gov.uk

IntroductionArticle 11 of the EU Water Framework Directive(2000/60/EC) (WFD) requires that in catchmentswhere the basic programme of measures are deemedto be insufficient for achieving the qualityobjectives, supplementary measures are to bedesigned and implemented. These measures shouldbe technically feasible, cost effective andenvironmentally sustainable. Supplementarymeasures for high status water bodies are ofparticular importance due to their sensitivity toanthropogenic impacts and the catchment specificnature of many threats. While alternative qualityobjectives may be set for waters in certain specifiedcircumstances, this does not apply to high statuswater bodies which may not be allowed todeteriorate.

The basic measure to control phosphorus (P) use inagriculture under the WFD is the Nitrate regulations(S.I. No. 610 of 2010). Frequently these regulationshave few consequences for more extensively farmedareas, which often coincide with the location of highstatus water bodies. However, despite low stockingdensities, phosphorus export in such areas can stillpose a threat to high status water bodies and mayrequire supplementary measures to mitigate Pexport effectively. Targeting critical sources areas(CSAs) of P in such catchments could significantlyimprove the environmental efficiency and costeffectiveness of supplementary measures whileminimising the impact on farm management. Thispaper considers how a CSA-approach to thedevelopment of P supplementary measures could beintegrated into the exiting WFD structures, throughan exploratory case study of the implementation ofsuch an approach in the Lough Melvin catchment.

MethodsLough Melvin is unique amongst Irish lakes,supporting a fish community typical of a naturalpost-glacial salmonid lake. Due to its uniqueecosystems and biodiversity it has been designatedas a Special Area of Conservation (SAC) under theEU Habitats Directive (92/43/EEC). In addition,

three sub-catchments are classified as high statuswater bodies under the WFD.

Between 1991 and 2007, P concentration in the lakeincreased from 19 to 29 µg L-1 bringing the nutrientstatus of the lake close to eutrophic. Despite thepredominance of extensive farming (averagestocking density of 0.5 LU/ha) agriculture isestimated to be contributing 62% of the P load toLough Melvin. Total P export from the Glenaniffsub-catchment, which has little forestry, a sparsepopulation and is classed as a high status waterbody, increased from 0.71 tonnes yr-1 in 1990 to1.27 tonnes yr-1 by 2007.

Figure 1: Development of P mitigation measures inthe Lough Melvin catchment (Doody et al. 2009)

Due to the increase in P export from agriculture,Schulte et al. (2009) implemented a study in thecatchment with the aim of developing technicallyfeasible, cost effective and environmentallysustainable mitigation measures. Throughout thestudy, an explicitly participatory approach wasadopted in which the input of local farmers andagricultural stakeholders were integrated into thedevelopment of the measures (Doody et al. 2009)(Figure 1). A field-by-field risk assessment, usingthe P risk index approach, was carried out on 50farms in the catchment in conjunction with semi-structured interviews with each farmer. This datawas then used to develop a preliminary list ofmeasures which were presented at a stakeholderworkshop. Subsequently the measures arising fromthe workshop were evaluated by the farmers and forcost effectiveness and a final list of measures to

mitigate P export from agriculture in the catchmentidentified.

ResultsThe field-by-field risk assessment highlighted that31% of the fields surveyed were a high risk for Ploss, with a further 30% categorised as a mediumrisk. In total 20 measures were identified to mitigatethe risk of P loss from these fields. The costeffectiveness and preference of the farmers for sixof the top measures is presented in Table 2. Themost cost effective measures were also those rankedmostly highly by farmers. Schulte et al. (2009)calculated that implementation of these measureswould result in a 36% reduction in the total Pexported for less than 1% of total potential costs;while over 50% of the total potential reduction in Ploss could be achieved at just over 5% of totalpotential costs. .

Table 2. The cost effectiveness and farmerpreferences for a selection of the P mitigationmeasures identified for the Lough Melvincatchment. Category A indicates highest cost-effectiveness/ highest popularity.

DiscussionArticle 14 of the WFD encourages the activeinvolvement of stakeholders in the development andimplementation of the WFD. As such the approachof Schulte et al. (2009) provides a template bywhich cost-effective scientifically robustsupplementary measures can be developed inconsultation with the local farming community. Tohelp achieve the objectives of the WFD, thisapproach could be integrated into a tiered riskassessment that builds on the WFD Article 5 riskassessments and uses (for example) the P IndicatorTool (Heathwaite et al., 2003) to identify high riskareas for P loss at a scale of 1 km2 withincatchments with high status water bodies. Theimplementation of the methods of Schulte et al

(2009) would then be focused within these 1 km2

high risk areas and supplementary measuresdeveloped from the outputs of the survey.

ConclusionsThe development of effective supplementarymeasures is vital to protecting the water quality andbiodiversity in high status water bodies. The threatsposed to many high status water bodies arecatchment specific and as such, supplementarymeasures should be developed that take the sitespecific nature of those threats into consideration.This paper has presented how such an approachcould be successfully utilised to identifysupplementary measures to mitigate P loss incatchments with high status water bodies.. Theparticipation of farmers and other stakeholders iscentral to the successful development of measuresthat are cost effective, environmentally efficient andthat can be successfully implemented withinexisting farming systems.

ReferencesDoody, D.G., Schulte, R.P.O., Byrne, P. and Carton,O.T. (2009). Stakeholder participation in thedevelopment of agri-environmental measures.Tearmann: Irish journal of agri-environmentalresearch 7: 229-241.

Heathwaite A. L., Fraser A. I., Johnes P. J.,Hutchins M., Lord E. and Butterfield D. (2003). Thephosphorus indicators tool: a simple model ofdiffuse P loss from agricultural land to water. SoilUse Management 19: 1–11

Schulte R.P.O., Doody D.G., Byrne P., Cockerill C.and Carton O.T. (2009). Lough Melvin: Developingcost-effective measures to prevent phosphorusenrichment of a unique aquatic habitat. Tearmann:Irish journal of agri-environmental research 7: 211-229.

Measures Costeffectiveness

Popularity

Feed low P conc. A A

Zero P on Index 4silage area

A B

Free advisory service A A

Reduce stocking rate(sheep)

B B

Sedimentation barriersin ditches

B A

Reduce stocking rate(suckler cows)

B B

Run-off interceptionditches

C B

HNV farming policy – progress to-date andfuture challenges

G. BeaufoyEuropean Forum on Nature Conservation andPastoralism. 24 Broad Street, Stratford-on-Avon, CV37 6HN, UKEmail: gbeaufoy@talktalk.net

IntroductionThe HNV farming concept emerges from arecognition that certain patterns of farmlandare inherently of high biological richness,especially when existing at a landscape scale,particularly landscapes that contain asignificant proportion of land in a semi-naturalcondition, such as permanent pastures andmeadows, grazed woodlands and traditionalorchards.

Many studies have shown that greaterheterogeneity, connectivity and area of semi-natural elements in an agricultural landscapetend to have a positive influence on speciesrichness and abundance, across a range ofwildlife groups. The semi-natural patches needto be not only of sufficient quality, but also ofsufficient size and connectivity. At a generallevel, experts in France have estimated that fora farming landscape to support high levels ofbiodiversity, it should include a minimum of20% semi-natural land cover (Le Roux et al.,2008).

In intensified agricultural landscapes, thesebeneficial conditions for biodiversity havebeen lost and their (partial) restorationgenerally occurs only at a cost to publicfinances, e.g. through agri-environmentschemes. On HNV farmland these conditionsbroadly exist already, and the biodiversitychallenge is to prevent their decline in the faceof powerful socio-economic and policydrivers.

Semi-natural farmland types, and theirassociated biodiversity, extend well beyonddesignated protected areas. According to EEAestimates, HNV farmland may account for 30per cent of EU farmland (Paracchini et al.,2008). If we want to conserve nature inEurope, we should look at this big land-usepicture, not just the detail of how to manageparticular habitats within designated areas.

Many of Europe’s semi-natural farmlandhabitats are continuing to decline in area,and/or are in poor condition. A main andcontinuing cause of this decline is a parallelprocess of intensification on land with moreproduction potential, and abandonment of landwith production limitations. In some regions,this process can be seen at the scale of entirelandscapes (e.g. in parts of southern andeastern Europe). In most of the lowlands ofnorth-west Europe, where semi-naturalfarmland no longer occurs at the landscapescale, similar processes of abandonment andintensification continue to affect the small-scale remnants, for example on semi-naturalfields or parts of fields within farms.

In response to these processes and in order tohalt biodiversity decline, there is an urgentneed to develop an economic basis for farmingon semi-natural land. This requires acombination of public remuneration for thepublic goods generated, through targeted CAPpayments, and the development of greatersocio-economic resilience, for examplethrough marketing of high-value farmproducts. Both need to occur on a scalecommensurate with the challenge existing onthe ground.

In simple terms, HNV farming is a policyconcept that aims to give greater priority tofarming landscapes that still retain asignificant proportion of semi-natural landwithin the farming system, whether within oroutwith designated areas such as Natura 2000sites. It promotes a more strategic andintegrated approach than occurs under presentpolicies, by taking greater account of thesocio-economic realities of different farmingsituations, and recognising that these realitiesare central to achieving biodiversity goals.

Since 2005, identifying, supporting andmonitoring HNV farmland and farmingsystems have been priorities for EU ruraldevelopment policy. The European Forum onNature Conservation and Pastoralism(EFNCP) has been closely involved in thedevelopment of suitable approaches to thesetasks at the European level. Exploring how itall should work, through real examples at locallevel, is a particularly important complementto top-down desk studies, and in 2010 EFNCPjoined up with local partners to run a series of

projects to examine various aspects of HNVfarming and policy in parts of England, Wales,Ireland, France and Spain (Navarra) (seehttp://www.efncp.org/projects/). Funding hasbeen from local/national partners and from DGEnvironment of the EC. These projects aim toexplore how HNV farmland and farmingsystems can be identified and their socio-economic needs assessed, as the basis fordeveloping strategies for their effective long-term support.

EU policy contextAlthough the concept was first introduced in1993 (Baldock et al., 1993), HNV farming didnot become firmly established in EU policyuntil the start of the present rural developmentprogramming period. In its 2006communication on halting biodiversitydecline, the European Commissionemphasised that Natura 2000 and theconservation of threatened species will not beviable in the long-term without a widerterrestrial, freshwater and marine environmentfavourable to biodiversity. Key actionshighlighted included optimising the use ofavailable measures under the reformed CAP,notably to prevent intensification orabandonment of high nature value farmland,woodland and forest and supporting theirrestoration (COM, 2006)

In the same year, the EAFRD regulationStrategic Guidelines on rural developmentestablished HNV farming as one of threepriorities for Axis 2 of Rural DevelopmentProgrammes (RDPs). In order to includeeffective measures for HNV farming in theirRDPs, Member States should assess thepractical needs on the ground and how best toaddress them. Specifically, the EAFRDimplementing regulation states that theyshould produce an analysis of:

“Environment and land management: thehandicaps facing farms in areas at risk ofabandonment and marginalisation; overalldescription of biodiversity with focus on thatlinked to agriculture and forestry, includinghigh nature value farming and forestry systems[…]”

The 2007-2013 RDPs should have measures inplace to maintain HNV farming and forestrysystems. The effects of programmes will be

evaluated against this objective, by applyingspecific HNV indicators under the CommonMonitoring and Evaluation Framework(CMEF).

The CMEF Result Indicators include “Areaunder successful land managementcontributing to: (a) biodiversity and highnature value farming/forestry (e) avoidance ofmarginalisation and land abandonment”.

The CMEF Impact Indicators include“Maintenance of high nature value farmlandand forestry”, for which a baseline situationmust first be established.

The European Evaluation Network for RuralDevelopment (EENRD) Help Desk hasproduced guidelines for the application ofHNV indicators (Beaufoy and Cooper, 2008).These are intended to help Member Statesassess the baseline situation of HNV farmingand monitor how it evolves.

In 2010, the Help Desk organised a ThematicWorking Group of experts to develop guidanceon the application of all CMEF indicators,including HNV indicators. The report of thisgroup adds further practical detail andexamples of current practice to the 2009guidance (Lukesch and Schuh, 2010).

Summarising the thinking in these documents,HNV farming can be seen from twoperspectives:

- HNV farmland: land-cover types suchas semi-natural pastures/meadows andtraditional orchards, mosaics of low-intensity crop types with semi-naturalpatches.

- HNV farming systems: how the aboveland-cover types are managed,especially grazing and mowingregimes, low-intensity systems of cropmanagement and weed control inorchards, practices such astranshumance. Socio-economicaspects are also part of the farmingsystem, for example income levels,part-time farming.

To be meaningful and effective, monitoringand evaluation should address both these

perspectives, so that policy makers designingprogrammes have the information they needconcerning trends in the extent and conditionof HNV farmland types, in the farmingsystems and practices that use them, and intheir socio-economic situation.

Progress to date and future challengesWhen the current round of RDPs wasintroduced in 2007, few if any Member Stateshad a reliable basis for estimating theirbaseline extent or condition of HNV farmland.Some turned to the JRC-EEA CORINE-basedHNV farmland maps and figures, which arenot intended for use as CMEF indicators, orused the extent of Natura 2000 sites as aproxy, or simply provided no baselineindicator due to lack of data.

Since then, a great deal of work has beendeveloped across the EU by national agenciesand research bodies. Much of the effort to-dateis focused on developing methods for theidentification of HNV farmland, using nationaldata sets. Work has moved on from the basicmapping of land-cover types and protectedareas as first introduced by EEA, to morecomplex combinations of criteria includingaspects such as field size and land-usediversity. Many Member States have workongoing that is generating robust results; thefollowing are a few examples.

In Spain, a project started by the Ministry ofEnvironment in 2008 to identify HNVfarmland is now in its second phase under thenew combined Ministry of Environment andRural and Marine Affairs. This work putsconsiderable emphasis on species and habitatdistribution as indicators of HNV farmland.Meanwhile the regional government ofNavarra has a project with EFNCPinvolvement to develop HNV farming andforestry indicators at the regional level, andalso sets of indicators for monitoring changesin HNV farming and forestry systems withinthe region.

A study in Finland is developing the nationalHNV farmland indicator based on 5-8 farm-level variables with varying weighting factors.The amount of semi-natural grasslands,permanent grasslands, grazing animals andarea in special agri-environment schemecontracts supporting biodiversity will have a

relatively large weight and some other farmingstatistics lower weights in determining theHNV value of a farm. The total area of farmsexceeding a threshold value will be consideredas HNV farmland, the area of which will bemonitored at intervals of a few years.

HNV farming is largely about determiningpriorities, and in this context, processes fordialogue are as important as desk studies. InEstonia a national HNV working group wasformed in July 2009 by the AgriculturalResearch Centre (ARC). This working groupincludes representatives from the Ministry ofAgriculture and Ministry of Environment,State agencies administrated by theseMinistries, and experts from EstonianUniversity of Life Sciences. The HNVworking group sees its work in three stages: 1)Identifying and working out HNV baselinecriteria; 2) Evaluation of the applicationpossibilities; 3) Proposals for HNV in the RDPcontext (e.g. support measures, combinationsof measures).

Much of the work undertaken by national andregional authorities has focused on land cover,and especially on methods of mapping theland-cover types associated with HNVfarming. This work has tended to come upagainst the limitations of European andnational data bases. CORINE provides only avery approximate picture, and national land-use data are often not much more useful.However, experience shows that thechallenges are far from insuperable.

The Land Parcel Identification System (LPIS)which all Member States are required todevelop for administering CAP paymentsoffers potential solutions. Some countries havemore detailed, and therefore more useful, LPISthan others. For example, LPIS in Spain hasnumerous categories of permanent pasture,some of which are by definition semi-natural,such as pastures with scrub and trees.Identification of broad types of HNV farmlandin Navarra (Spain) has been possible usingLPIS in combination with regional land usedata bases. In England by contrast, LPIS hasonly one category of permanent pasture, thatincludes a range of types from moorland tohighly productive pastures that are frequentlyreseeded, and is thus less useful.

LPIS becomes even more useful when data onthe presence of environmentally valuable landcover is recorded on the system. This is thecase in Slovakia, where a national survey ofsemi-natural grasslands has been undertakenand incorporated in LPIS, thus producing anaccurate and detailed baseline for the maintype of HNV farmland in the country.

As explained in another conference paper (seepaper by D. Fuchs), Germany has established abaseline and monitoring system for HNVfarmland extent and condition, using asampling approach similar to that used bymany countries for the Farmland Birds Index,and also by the UK Countryside Survey.

Farming systems have received less attention.In Italy, data from the Farm Structures Surveyhas been used to estimate the distribution andextent of broad farm types with characteristicssuch as low livestock density. The governmentproject in Navarra has developed a typology ofbroad HNV farming and forestry systems, andis studying these to establish indicators thatfocus on aspects such as particular farmingpractices characteristic of the system andknown to be vulnerable to change. Indicatorsof the socio-economic condition of HNVfarming systems will also be considered, inorder to allow effective policy responses to thethreat of abandonment. As with HNVfarmland condition, a sampling approach isprobably the most practical way of monitoringthese more detailed aspects of HNV farmingsystems.

Turning to the EAFRD priority for supportingHNV farming, this has been taken up mostexplicitly by new Member States, with manygiving considerable emphasis to the theme intheir RDPs. The Romanian and BulgarianRDPs mention HNV farming 58 and 82 timesrespectively, compared with 4 times in theRDP for Wales. Some ambitious schemes havebeen established, such as the agri-environmentmeasure for HNV grassland in Romaniatargeting the Carpathian uplands zone, andmeasures for HNV orchards in Bulgaria.Slovakia has developed a country-widemeasure for all HNV grasslands registered onLPIS.The old Member States have been moreinclined to assume that what they are doingalready ticks the “HNV box”, especially

through agri-environment. However, in a localHNV farming project in south-west Englandwe found that considerable areas of HNVfarmland were not participating in agri-environment measures, suggesting a need toreview current approaches.

ConclusionsEFNCP believes that basic economic supportfor maintenance of HNV farmland is besttargeted through farm-level criteria. At itssimplest, this would consist of a premiumpayment for specific land-cover typesidentified on LPIS, particularly permanentpastures and meadows, in return for aminimum level of positive management and acommitment not to intensify throughreseeding.

This would provide an incentive for preventingabandonment and intensification and thus formaintaining the basic public goods values ofthese land-cover types, right across the EU.Such a measure is perfectly feasible at presentwith no need for new maps or zoning.

A more sophisticated approach to bedeveloped over time is to record on LPIS morespecific land-cover types, as has been done inSlovakia with the HNV grassland survey. Thiswould allow implementation of a premium forother semi-natural farmland types, such astraditional orchards.

In parallel, simple systems for monitoring theextent and condition of these HNV farmlandtypes can be set up, using LPIS withappropriate categories (Spain), baselineinventories as in Slovakia, Estonia and Wales,and sample surveys as in Germany.

HNV farming systems require a different butcomplementary approach. The support needsare various and quite complex, including theneed to maintain and incentivise particularfarming practices and combinations ofpractices, to improve socio-economic viabilityand living conditions of farming communities,and to achieve a critical mass at the landscapescale by encouraging farmers to work together.In terms of policy architecture, Pillar 2 seemsthe appropriate place for putting togetherprogrammes of measures to address theseneeds.

Programmes should target particular HNVfarming systems, for example those ofexceptional public goods value, and/or that arehighly threatened, as has been done in severalMember States already. Indicative maps ofHNV farming systems may be useful for thispurpose, as illustrated by the work in Navarra,not necessarily to define precise boundaries ofzones eligible for support, but as a way ofprioritising systems and structuring tailoredsupport programmes.

Policy makers have been known to declare thatthe HNV farming concept is unclear, orimpossible to implement. Experience is nowshowing that this is not the case, especially ifthe concept is broken down into manageablecomponents – distinguishing measures tosupport HNV farming from instruments formonitoring and evaluation; distinguishingsemi-natural farmland from other types offarmland; and distinguishing HNV farmlandfrom HNV farming systems.

There are data difficulties, but these can beovercome at no great cost, as has been shown.The result should be more effective policydesign and more cost-efficient delivery of EUgoals.

We have failed to meet our goals for haltingbiodiversity decline by 2010, and are nowsetting new ambitious targets. To meet these,we need an integrated system for monitoringnot only biodiversity itself, but also the typesof farmland and farming that are of mostbiodiversity value. This should be done in away that generates meaningful information forpolicy makers, leading to policy adaptationand improvement.

ReferencesBaldock, D., Beaufoy, G., Bennett, G. andClark, J. (1993). Nature Conservation andNew Directions in the Common AgriculturalPolicy. Institute for European EnvironmentalPolicy, London.

Beaufoy G and Cooper T, 2008. GuidanceDocument to the Member States on theApplication of the HNV Impact Indicator.EENRD Help Desk

COM (2006) 216 Final Communication fromthe Commission halting the loss of

biodiversity by 2010 — and beyond.Sustaining ecosystem services for humanwell–being.

Le Roux X., Barbault R., Baudry J., F. Burel,F., Doussan, I., Garnier E., Herzog F., LavorelS., Lifran R., Roger-Estrade J., Sarthou J. P.and Trommetter M.(eds) (2008). Agriculture etbiodiversité. Valoriser les synergies. Expertisescientifique collective, synthèse du rapport,INRA (France).

Lukesch, R. and Schuh, B. 2010. Workingpaper on approaches for assessing the impactsof the Rural Development Programmes in thecontext of multiple intervening factors.Findings of a Thematic Working Groupestablished and coordinated by The EuropeanEvaluation Network for Rural Development.

Paracchini M.L, Petersen J-E., Hoogeveen Y.,Bamps C., Burfield I., van Swaay C. (2008).High Nature Value Farmland in Europe. Anestimate of the distribution patterns on thebasis of land cover and biodiversity data.European Commission Joint Research Centre,Ispra.

Evidence on the environmental effectiveness ofREPS: a literature review

J.A. Finn and D. Ó hUallacháin1Teagasc, Environment Research Centre,Johnstown Castle, WexfordEmail: john.finn@teagasc.ie

IntroductionAs a formal requirement of the Rural DevelopmentRegulation, Member States are obliged to monitorand evaluate the environmental, agricultural andsocio-economic impacts of their agri-environmentprogrammes. Summary reports on policyevaluation of agri-environment schemes haveconcluded that there has been insufficientmeasurement of their precise environmentaloutcomes (DG Agriculture 2004; Oréade-Brèche2005). The European Court of Auditors will reportin 2011 on its audit of the environmentaleffectiveness of agri-environment schemes, whichmay lead to significant changes in policyimplementation. The World Trade Organisation(WTO) also requires that the environmentalbenefits of agri-environmental payments areclearly demonstrated, to ensure that such paymentsare not disguised trade subsidies. One of the best(if not only) ways to address these variouspressures is to quantitatively demonstrate theenvironmental benefits and value-for-money ofagri-environment schemes. This policy contexthighlights the need for quantitative demonstrationof the environmental impact of agri-environmentschemes, and why this will become increasinglyimportant.

A variety of research projects have beenundertaken that investigate the environmentaleffects of REPS, through an examination of eitherspecific environmental measures or specificgeographical areas. Here, we review thepublications from these projects, and assess theextents to which they can inform an environmentalassessment of REPS.

Materials and MethodsAn attempted systematic review with a number ofvarious relevant search terms in Web ofKnowledge only resulted in a total of about tenrelevant research articles. As an alternative, wecollated and reviewed available literature on thesestudies, with an emphasis on empirical researchthat is directly relevant to the environmentaleffects of REPS.

Results and DiscussionA variety of research projects have been conductedon REPS. Publications from these are groupedunder the relevant broad environmental objectives

as indicated in Table 1, and each of these groupsdiscussed in turn. This list is not intended to beexhaustive, but includes most of the publishedresearch studies as well as many of theunpublished ones.An increasing number of studies are available withwhich to learn about the actual or likelyenvironmental effectiveness of REPS. Aconsiderable proportion of these studies has notbeen published in international journals, and isonly available as national reports, theses,conference papers and conference abstracts.Compared to the high standard of evidenceassociated with journal articles, care is required inthe interpretation of evidence from other sources(although some of this is of a very high standard).On the basis of this review, a number ofconclusions arise that are relevant to institutionalefforts to assess the environmental impacts ofREPS: There is insufficient evidence with which to

judge the environmental effectiveness of thenational-scale implementation of the wholeRural Environment Protection Scheme. Thismakes it equally likely that the full benefits ofthe scheme have not been measured, as well asreducing the opportunity to learn how toimprove it.

Some individual studies provide evidence toscientifically assess the environmental effect ofindividual REPS measures; however, moststudies lacked national-scale coverage.

There is a distinct lack of studies that usebaseline data to compare change over time(longitudinal studies).

Of the studies undertaken to date, there hasbeen an emphasis on biodiversity studies, butthese have had little or no co-ordination intheir aims, methods, temporal scales or spatialscales.

There have been surprisingly few studies onthe impact of REPS on nutrient managementand water quality, but the available evidence isgenerally positive.

A considerable number of studies haveinvestigated the environmental effects ofREPS, although relatively few of these havebeen published in journals.

Some evidence currently exists to guideadvice/recommendations about theenvironmental effectiveness of REPS.

ConclusionsA review of available publications confirmed theabsence of a comprehensive, national-scale studyof the environmental impacts of REPS. Because ofthis, there is insufficient evidence with which tojudge the environmental effectiveness of thenational-scale implementation of the whole

scheme. It is important to note that this does notnecessarily mean that REPS has not deliveredenvironmental benefits, but that there has beeninsufficient collection of evidence of theenvironmental performance of the whole REPSprogramme. Thus, the full benefits of the schemehave not been measured, and there has beenreduced opportunity to learn how to improve thescheme. For many of the newer supplementarymeasures and options that have been introducedsince REPS 3, no empirical evidence is availablewith which to judge their ex post environmentaleffects, which hinders an overall assessment of thewhole scheme. For some specific measures,however, sufficient evidence is available to informan objective assessment in some cases, and to helplearn how to improve environmental effectivenessin most cases.

AcknowledgmentsThis project was supported by Teagasc RMIS6024.

ReferencesDG Agriculture (2004) Impact assessment of ruraldevelopment programmes in view of post-2006rural development policy. Brussels. EuropeanCommission.

Finn, J.A. and Ó hUallacháin, D. A review ofevidence on the environmental impact of Ireland’sRural Environment Protection Scheme (REPS).Biology and Environment (in press).

Oréade-Brèche (2005) Evaluation of agri-environmental measures. Annex 14: Etudenationale Irlande. Auzeville. Oréade-Brèche.

Table 1. References relevant to the environmental effectiveness of REPS across different environmentalcategories. (Full list of references available in Finn and Ó hUallacháin, in press.)Topic ReferencesNutrient management andgaseous emissions

McEvoy (1999), Casey and Holden (2005, 2006), Lanigan et al. (2008), Hyneset al. (2007, 2008b), Richards et al. (2007), Doody et al. (2009), Schulte et al.(2009)

Archaeology O'Sullivan (1998, 2001), Sullivan (2005, 2006), Sullivan et al. (1999)Measure A farmlandhabitats

Dunford and Feehan (2001), Williams et al. (2009), Walsh (2009), Visser et al.(2007), Moran et al. (2008), NPWS (2008), van Rensburg et al. (2009),O’Rourke and Kramm (2009)

Non-designated farmlandhabitats

Hickie et al. (1999), Bohnsack and Carrucan (1999), DAF (1999), Jones et al.(2003), Hyde (2003), Aughney and Gormally (2002), Gabbett and Finn (2005),Copland (2009), Copland and O’Halloran (2010), Egan (2006), Hynes et al.(2008a), Speight (2008), Purvis et al. (2009a, p. 17-20)

Field margins Feehan et al. (2005), Fritch et al. (2009, 2011), Purvis et al. (2009a), Sheridanet al. (2008, 2009)

Hedgerows Flynn (2002), Copland et al. (2005), Copland (2009)Assessment acrossmultiple environmentalobjectives

Bartolini et al. (2005), Viaggi et al. (in press), Finn et al. (2007, 2009), Finn etal. (2008b), Kelly (2008), Carlin et al. (2010),

Financial effects McEvoy (1999), Connolly (2005), Connolly et al. (2005, 2006, 2009), Kinsellaet al. (2007ab) (and others)

Others In addition to the above, a number of other publications address a variety ofissues in the context of REPS, including landscape preferences, economiccommentaries and general critiques. These are listed in Finn and Ó hUallacháin(in press).

Invasive species control Case Study: Hottentotfig (Carpobrotus edulis) control in Ireland.

N. Smyth 1, M. Jebb1, and A. Booth2

1National Botanic Gardens, Glasnevin, Dublin 9,2Conservation Services, Kilmessan, Co. MeathEmail: noeleen.smyth@opw.ie

IntroductionHottentot fig (Carpobrotus edulis) is a populargarden plant from South Africa. Unfortunately it isalso an aggressive invader of coastal habitats,forming vast mats to the exclusion of all otherplants. On the Gower peninsula of Wales andalong the Cornish and Devon coasts of GreatBritain it has formed extensive coloniessmothering many kilometres of coastal cliffs. Onthe drier eastern coasts of Ireland it poses a seriousecological threat.

Plate 1. Hottentot Fig.

Nothing eats Hottentot fig in Ireland, thus a colonynot only displaces native plants, which are a foodsource for bees, butterflies and moths, but thedense carpets represent a dead zone in regard toinsects, and thus the birds that feed on them (theonly known beneficiary are rats, that eat the fruits).The consequence is that the cliffs are becoming alife-less zone for insects and birds. This projectaimed to ultimately eradicate Hottentot fig fromHowth Head, Co. Dublin with initial pilot studiescarried out in The Murroughs, Co.Wicklow andcontrol efforts are due to be expanded to all knownIrish sites (Reynolds, 2002) in 2011.

Materials and MethodsPilot control experimentTo determine the best control method to use forHottentot fig, herbicide experiments were carriedout in Wicklow during March 2010. Physicalcontrol was deemed fruitless, due to the issues ofdealing with the hazardous waste and the sheervolume of material (stems were found to weight inexcess of 10kg). The area affected by Hottentot figon Howth Head lies within Howth Head SAC000202, so no experimental work or treatmentscould be carried out during nesting season March -September. The pilot experimental chemicaltreatments were carried out at an alternative coastalsite in Wicklow.

Tetrads of 1.5m2 (2.25 m2) were treated with eachof the three herbicide products trialled. The patcheswere marked using white pebbles, and each patchwas separated by 1m from the next treatment. Thespray was applied until all leaf surfaces were fullywetted. About ¼ litre was used for the two tetradsof each treatment.

Herbicides used in pilot experiment1. B&Q Lawn weedkiller (0.358 g/l mecoprop-Pand 0.191 g/l dichlorprop-P soluble).2. Resolva weedkiller 24H Action Westland gardenhealth (3 g/l glyphosate and 0.3g/l diquat).3. Monsanto, Fast action Roundup weedkiller (7.2g/l glyphosate acid, present as 9.7 g/lisopropylamine salt of glyphosate).

The treatments were applied on a dry still day. Norain fell in the following 24 hours, and theherbicide was applied according to themanufacturer’s instructions to preventcontamination of the surrounding habitat. Afterthirty seven days, results from the pilot experimentwere obvious, using 3 g/l glyphosate and 0.3g/ldiquat (Resolva Weedkiller 24 H Action) theHottentot fig mortality rate was >95%.

Plate 2. Chemical treatment (3 g/l glyphosate and0.3g/l diquat) of Hottentot fig showing >95%mortality.

Control of Hottentot Fig on Howth HeadThe results of the pilot study helped demonstratethe most effective herbicide. The target site atHowth head was tackled during October 2010during dry calm days. A wheelbarrow powersprayer (KS, 120 litre tank, petrol motor) waswheeled to access points along the cliff path andthe 30m hose extension from the tank meant thatoperators could access the Hottentot fig along thecliff without having to wear a cumbersomeknapsack sprayers. Knapsack sprayers (10 litre)were used at the sites near Sutton Sailing Club, Seacottage and Lion’s Head where it was safe foroperators. Volunteer labour was used in the mainand professional climbers were used where thecliff face was too steep for volunteers. Thechemical found to be most successful in the

Wicklow trails (3 g/l glyphosate and 0.3g/l diquat)was mixed on site. Water was filled at Howth pieror brought along in drums and transported bywheelbarrow once on site. The site was revisited inFebruary 2011 and chemical was applied to smallpopulations which were missed previously. Apreliminary botanical survey of speciesregeneration at treated sites was also carried outusing 50cmx50cm quadrats, placed whereHottentot fig had been treated. Species presentwere identified where possible using Webb,Parnell & Doogue (1996).

Results and DiscussionBest control results were obtained using 3 g/lglyphosate and 0.3g/l diquat (Resolva 24 H) withmortality rate >95%. Seed pods treated withchemical were collected from dead and dried outstems in spring 2011 and the seed contained withinwere found to be capable of germinating andgrowing. This was a worrying result as there wasevidence at many of the sites of gnawing of theseed pods by rats.A survey in spring 2011 of the emerging likelyreplacement vegetation to Hottentot fig, found thatnative grass species such as Festuca rubra,Agrostris stolonifera & Dactylis glomerata weredominant replacement species, with seedling ofherbs such as Cochlearia sp., Armeria maritima,Plantago maritima and Apium graveolens.

ConclusionsThe herbicide treatment that worked mosteffectively on Hottentot fig was 3 g/l glyphosateand 0.3g/l Diquat (Resolva 24H Action) which wasused to treat the invaded sites at Howth Head.Other sites which were discovered duringfieldwork near Sutton and Dun Laoghaire werealso included in herbicide control in late October2010. The herbicides used in this control effortdiffered to those recommended by Kelly &Maguire (2009) for Hottentot fig control. The pilotchemical trails in Wicklow clearly demonstratedthat the ratio mixture of glyphosate and diquat wasthe more effective as a “killing cocktail”.Following the success of the project in Howth, wewill expand and treat Hottentot fig at Carlingford,Wexford, Waterford and Cork i.e. all the currentlyknown locations of this invasive species in Ireland(Reynolds, 2002). This is the first project toaddress a known invasive alien species with asmall Irish range. We deem the policy of dealingwith known alien invasive species which are stillbelow “the radar” can provide that vital “stitch intime” before species such as Hottentot fig becomedominant on coastal cliffs throughout the Island.

AcknowledgmentsThe project team wish to thank the HeritageCouncil for providing funds to carry out the controlproject. Fingal Co. Council distributed informationposters and leaflets to all the local libraries andcontact was made with the local and national press.Some landowners were very generous with accessto invaded sites and provision of water for thechemical sprayers. We would like to thank inparticular Ann Murphy, Hans Visser, DeborahTierney and Florence White of Fingal Co. Council,The members of the Howth Committee, MauriceEakin and Niall Harmey of NPWS. The Baileylighthouse keeper Brendan and Captain Ciaran OHiggins of Irish Lights for granting access.

ReferencesReynolds, S.C.P. (2002) A catalogue of alienplants in Ireland. Pp 67. National Botanic Gardens,Glasnevin.

Webb, D.A., Parnell, J. & Doogue, D. (1996) AnIrish Flora. Dundalgan Press (W. Tempest) LTD.Dundalk.

Kelly, J. and Maguire, C.M. (2009). Hottentot Fig(Carpobrotus edulis) Invasive Species Action Plan.Prepared for NIEA and NPWS as part of InvasiveSpecies Ireland.

Options for enhancing the biodiversity value ofintensive livestock farms: experience from theEnglish agri-environment schemes

S.R. MortimerCentre for Agri-Environmental Research,School of Agriculture, Policy & Development,University of Reading, Reading, RG6 6AR, UKEmail: s.r.mortimer@reading.ac.uk

IntroductionPermanent grassland makes up a greater proportionof the agricultural area in the UK and Ireland thanin any other EU country, representing 60% and72% of UAA respectively (Eurostat, 2007). Of thepermanent grassland in the UK, approximately half(about 6 million hectares) comprises improvedgrassland on moist or free-draining neutral soilstypical of lowland livestock farms. These swardstend to have low plant species richness and aretypically dominated by perennial ryegrass (Loliumperenne). The aim of this paper is to review theways in which biodiversity of such farmland canbe enhanced, focussing on the evidence behindmanagement options in English agri-environmentschemes (AES) at a range of scales and utilising arange of mechanisms.

Development of AES in EnglandAES were first introduced in Europe in the mid1980s, and have been obligatory for EU memberstates since 1992. In England, the initial focus wason protecting areas of high landscape valuecontaining threatened habitats of conservationimportance. This led to the designation ofEnvironmentally Sensitive Areas (ESAs) from1987, within which payments were available toencourage appropriate land management practices.The approach was broadened to the wider farmedlandscape in 1991 with the introduction of theCountryside Stewardship Scheme (CSS), whichsought to enhance the environment of areas ofcountryside outside of the ESAs. Within thisscheme, funds were still largely focussed onprotecting, restoring and creating high nature valuefarmland habitats. Concern about farmland birdsand other taxa typical of arable farmland led todevelopment of arable options under the ArableStewardship Scheme and these were incorporatedinto CSS from 2001. The Curry Report (PolicyCommission 2002) on the future of farming,advocated the introduction of a ‘broad andshallow’ scheme as a means of incentivisingenvironmental management amongst a largerproportion of farmers, and this led to thedevelopment of the current EnvironmentalStewardship scheme in England, comprising an‘Entry Level’ (ELS) and a ‘Higher Level’ (HLS)available from 2005. However, concerns have been

raised about the additionality achieved by the ELS,the weak spatial targeting at both landscape andwithin-farm scales, and the attractiveness ofoptions for intensive livestock enterprises (Hodgeand Reader 2010; Butler et al., 2007; Defra, 2008).

AES and grassland diversityThe botanical diversity of grasslands is primarilycontrolled by levels of soil fertility and disturbanceand their effects on competitive interactions. Inintensively-managed agricultural grasslands, highproductivity is promoted by inputs of inorganicfertilizers and the sowing of competitive grassvarieties, allowing higher stocking densities ormore frequent cutting of fields for silage. Thedecline in plant species richness and thereplacement of plant species of conservation valuewith those typical of eutrophic, disturbedconditions is further exacerbated by the use ofherbicides.

The invertebrate diversity of agriculturalgrasslands is driven by botanical composition andvegetation structure and their combined effects onfood resource abundance and microclimate. Theuse of insecticides, fungicides and veterinarypharmaceuticals limit invertebrate abundance anddiversity. The relationship between managementand the abundance of vertebrate taxa is morecomplex, relating not only to provision of foodresources, but also their accessibility andinteractions with factors such as predation risk.

Enhancement of grassland diversity under AESIn the early AES schemes in England, the focuswas on the protection of high nature valuegrasslands through prescriptions controllingstocking densities and input regimes (Critchley etal., 2004). As the focus of AES broadened toinclude restoration and enhancement, the problemsof high soil fertility and a paucity of sources ofpropagules or colonists in the landscape came tothe fore (Edwards et al., 2007; Woodcock et al.,2008, 2010). Consequently, grassland restorationoptions are targeted on sites of low fertility and/orclose to areas of high nature value grassland.

Given this, what is the biodiversity value of theapproximately 6 million hectares of agriculturallyimproved, species poor grassland in the UK? Inspite of the general negative impact of grasslandintensification on wildlife, some taxa do utiliseintensively-managed grasslands, for examplecertain bird species (e.g. starlings, swallows)(Møller, 2001). Improved grassland left to seed canprovide winter food resources for buntings andfinches (Buckingham and Peach, 2006). Highstocking densities can be useful in providing

sufficient dung fauna for bats (Jennings andPocock, 2009).

Within the ELS scheme, options for intensivelivestock farms can be grouped according to theirlocation within the landscape: (a) in-field options;(b) options for grassland field margins; (c) optionsfor boundary features; (d) options to protectadjacent habitats and (e) options to enhanceprovision of food resources in the wider landscape(Table 1). Of these options, those for hedgerowmanagement, ditch management and reducedinputs on permanent grassland are most frequentlychosen (Defra, 2008).

Table 1. Selected management options relevant tointensive lowland livestock farms (NaturalEngland, 2010).Code DescriptionManagement within grassland fieldsEK2-3 Permanent grassland with low inputsEK5 Mixed stockingEC2 Protection of in-field trees on grasslandEE7 Buffering in-field ponds in grasslandManaging margins of grassland fieldsEE4 2 m buffer strips on intensive grasslandEE5 4 m buffer strips on intensive grasslandEE6 6 m buffer strips on intensive grasslandEC25 Hedgerow tree buffer strips on grasslandManaging boundary featuresEB1-3 Hedgerow managementEB6-7 Ditch managementEB11 Stone wall protection and maintenanceEB12 Earth bank managementEC23 Establishment of hedgerow treesProtection of adjacent habitatsEC3 Maintenance of woodland fencesEC4 Management of woodland edgesEJ11 Maintenance of watercourse fencingIncreasing landscape-level diversityEF2 Wild bird seed mixtureEF4 Nectar flower mixtureEG4 Cereals for whole-crop silage

Development of options and evidence baseThe evidence base for the development of schemeprescriptions has a long history of funding byDefra (previously MAFF), especially throughBD14 programme. The focus has shifted from aninitial focus on high nature value grassland and therelationship between soil fertility, grazingmanagement and botanical composition, to a focuson wider grassland biodiversity with a strongemphasis on interactions between taxa in differenttrophic groups and providing a mechanisticunderstanding of factors controlling diversity andspecies composition.

Management within grassland fieldsOptions prescribing grazing or cutting regimes andinput use in grasslands may yield environmentalbenefits through the reduced management intensityprotecting natural resources, but gains inbiodiversity may be slow. Of the improved andsemi-improved grasslands sampled in the UK ESAschemes, only 30% showed signs of restoration ofbotanical diversity following the introduction ofprescriptions relating to fertilizer inputs andgrazing intensity (Critchley et al., 2004).

Some attempt has been made to identify plantspecies that have wildlife value and are likely topersist in the fertile conditions of improvedgrasslands (Mortimer et al., 2006). Many of theplant species identified had beneficial agronomiccharacteristics, including high productivity andfeed value, especially amongst the grasses andlegumes. However, most of the forb speciesidentified have problems persisting in swards onsoils of high fertility, although a number of robustspecies of high wildlife value were identified thatperform reasonably well in grassland enhancementschemes on moderately fertile soils, for exampleyarrow (Achillea millefolium), knapweed(Centaurea nigra), ox-eye daisy (Leucanthemumvulgare) and ribwort plantain (Plantagolanceolata). Options for the establishment ofwildflowers in grasslands are available in the HLSscheme, but not in ELS.

Management of grassland field marginsAs with arable fields, the biodiversity value ofgrassland field margins tends to be greater at themargins of fields rather than in the centre.Focussing agri-environmental management at themargins has benefits in terms of the protection ofboundary features and adjacent habitats, and isusually more attractive a strategy to farmers.However, the tendency for farmers to select agri-environmental options relating to field boundariesand field margins within the ELS may mean thatsome biodiversity objectives are not met (Butler etal., 2007).

Research on management options for the marginsof intensively-managed grassland, has been carriedout on various combinations of restrictions oninput and disturbance regimes. A recent study insouthwest England manipulated the swardarchitectural complexity and botanical compositionin intensive grassland field margins, throughvarious combinations of fertilizer rate, cattlegrazing and differences in the timing and height ofcutting. The study found responses differedbetween different insect groups, with beetles andbutterflies benefitting from the low input/lowdisturbance combinations (Woodcock et al., 2007),

but bumblebees only responding to the treatmentsin which botanical composition was manipulatedby seed sowing (Potts et al., 2009). In Ireland,more interventionist management, including theexclusion of livestock from field margins usingfencing, and the sowing of species-rich seedmixtures, has shown the potential of more costlyprescriptions (Sheridan et al., 2008).

Current uptake of field margin options in the ELSscheme (‘buffer strips on intensive grassland’) ismuch lower than that for the arable equivalents(Boatman et al., 2010), with the majority of pointsbeing earned for field boundary management ratherthan the creation of field margin buffer strips.

Management of boundary featuresOptions for management of hedges and ditches areamongst the most frequently chosen by farmers inELS agreements and typically make up one third ofthe points requirement (Defra, 2008). Whilst thebiodiversity benefits of sensitive management ofhedgerows and ditches are well documented, theenhancement of wet habitats on intensive farmscould be developed in the current schemes (Fig. 1).The creation of small scale wet features infarmland could not only provide new habitats forplants and animals, but also confer additionalbenefits in terms of regulation of diffuse pollutionand flooding (Bradbury and Kirby, 2006).

Fig. 1. In an experimental study, artificial bundswere created in farmland ditches in a mixedfarming area of Leicestershire.

In an experimental study, artificial bunds werecreated in farmland ditches in a mixed farmingarea of Leicestershire. Bunds were created inditches bordering permanent pasture and arablefields and compared to paired stretches of ditcheswithout bunds. The abundance of emergent aquaticinsects was significantly greater in bunded sectionsof ditch than in the controls, and significantlyhigher in the ditches adjacent to pasture than thosein arable fields (Aquilina et al., 2007). Low costbunding of existing ditches in pastoral areas

therefore has the potential to deliver biodiversitybenefits and contribute to other environmentalbenefits (Bradbury and Kirby, 2006).

Enhancing landscape level diversityAs a result of increased specialisation of farmenterprises and the loss of mixed farms, manylandscapes in the west of England are dominatedby intensively managed grassland or forage maize.Such areas have suffered particular declines infarmland bird populations, especially seed–eatingspecies. Intensively-managed grasslands andforage maize crops are poor sources of food forseed- and invertebrate-feeding birds and can behostile nesting environments. Uptake of options forthe provision of food resources for birds orinvertebrates, such as the wild bird seed mix or thenectar flower mix has proved to be low amongstlivestock farmers in the ELS scheme (Boatman etal., 2010).

Research by the RSPB, CAER and Harper AdamsUniversity College examined the potential ofcereal-based whole crop silage as an alternative tograss or maize silage production. Seed-feedingbirds were found to feed preferentially on fieldsgrowing cereals for whole crop silage in thesummer rather than on maize of grass silage fields,and also preferred barley stubbles over winter(Peach et al., 2011). The pattern of bird usage offields in summer and winter reflected differencesin the densities of seed-bearing plants. The costs ofproducing whole crop silage from cereals weresimilar to those for making maize silage, and lowerthan those for grass silage. Cereals grown forwhole crop silage and followed by overwinteredstubble are now an option in the ELS scheme.

Future prospectsIt is clear that the range of options available forintensive grassland farmers in the current schemesprovide the basic ingredients for enhancing thebiodiversity value of their land. However, manyresearch studies have underlined the importance ofspatial targeting of appropriate options, both atwithin-farm and landscape levels, if biodiversitygains are to be optimised. Therefore, the balance ofresources between ‘broad and shallow’ schemessuch as the ELS, with low administration costs andlittle guidance to farms regarding option choice,and more demanding prescriptions such as those ofthe HLS, needs careful examination.

Given the increasing focus on natural resourceprotection, for example the implementation of theWater Framework Directive, it is likely that someoptions within current AES may be incorporatedinto other policy types, such as cross complianceor regulation. For the more demanding options that

may remain in AES, there is still a considerableneed to investigate landscape-level factors, such asthe optimum density and spatial configuration ofagri-environmental measures, in order to informspatial targeting and the provision of appropriateadvice about option selection and placement.

Recent developments in AES implementation inother European countries, such as agri-environment agreements entered into by groups offarms in environmental co-operatives, offers amechanism which might provide greatereffectiveness for AES through the adoption of alandscape-level approach.

The current basis of payment levels in use forAES, that of ‘income foregone’, also needsrevision, as it takes no account of theenvironmental value of the land concerned. Assome land may now have been under AESagreement for approaching 25 years, the value ofthe habitat produced as a result of the schemeneeds to be recognised in the payment levelsreceived, especially if the long-term managementnecessary for delivery of biodiversity is to beincentivised. Schemes which incorporate a‘payment by results’ incentive to grassland agri-environmental measures are likely to contribute tothe development of the long-term approachnecessary for biodiversity enhancement (Klimek etal., 2008).

For those farmers without high nature value habitaton their land, it is important that we move towardsa culture where farmland biodiversity is notperceived by farmers to be a separate product, paidfor by the public at the behest of powerful NGOlobby, and that we seek to internalised wildlifefriendly farming as part of a sustainable productionsystem. There is a clear trend towards recognitionof the ability of certain agri-environmentalmeasures to deliver benefits across a range ofecosystem services and the research described hereillustrates the potential for such a multifunctionalapproach.

AcknowledgmentsThe work summarises findings of a number ofresearch projects funded by Defra and EnglishNature. Thanks are due to colleagues at CAER andpartner research organisations including the RSPB,IGER North Wyke, Harper Adams UniversityCollege, the Pond Conservation Trust and theAllerton Research and Education Trust.

ReferencesAquilina, R., Williams, P., Nicolet, P., Stoate, C.and Bradbury, R. (2007). Effect of wetting up

ditches on emerging insect numbers. Asp. Appl.Biol. 81: 261-262.

Boatman, N., Willis, K., Garod G. And Powe, N.(2010). Estimating the wildlife and landscapebenefits of Environmental Stewardship. FinalReport to Defra. London: Defra.

Bradbury, R.B. and Kirby, W.B. (2006). Farmlandbirds and resource protection in the UK: Cross-cutting solutions for multi-functional farming?Biol. Cons. 129: 530-542.

Buckingham, D.L. and Peach, W.J. (2006).Leaving final-cut silage in situ overwinter as a seedresource for declining farmland birds. Biodiv. &Cons. 15: 3827-3845.

Butler, S.J., Vickery, J.A. and Norris, K. (2007)Farmland biodiversity and the footprint ofagriculture. Science, 315, 381-384.

Critchley, C.N.R., Burke, M.J.W and Stevens, D.P.(2004). Conservation of lowland semi-naturalgrasslands in the UK: a review of botanicalmonitoring results from agri-environment schemes.Biol. Cons. 115: 263-278.

Defra (2008). Environmental Stewardship Reviewof Progress. London: Defra.

Edwards, A.R., Mortimer, S.R., Lawson, C.S.,Westbury, D.B., Harris, S.J., Woodcock, B.A andBrown, V.K. (2007). Hay strewing, brushharvesting of seed and soil disturbance as tools forthe enhancement of botanical diversity ingrasslands. Biol. Cons. 134: 372-382.

Eurostat (2007). Agricultural statistics Data 1995-2005. Luxembourg: Office for OfficialPublications of the European Communities.

Hodge, I. and Reader, M. (2010). The introductionof Entry Level Stewardship in England: Extensionof dilution of agri-environmental policy? Land UsePol. 27: 270-282.

Jennings, N. and Pocock, M.J.O. (2009).Relationships between sensitivity to agriculturalintensification and ecological traits ofinsectivorous mammals and arthropods. Cons.Biol. 23: 1195-1203.

Klimek, S., Kemmermann, A.R., Steinmann, H.H.,Freese, J. And Ijsselstein, J. (2008). Rewardingfarmers for delivering vascular plant diversity inmanaged grasslands: a transdisciplinary case studyapproach. Biol. Cons. 141: 2888-2897.

Møller, A. P. 2001. The effect of dairy farming onbarn swallow Hirundo rustica abundance,distribution and reproduction. J. Appl. Ecol. 38:378-389.

Mortimer, S.R., Kessock-Philip, R., Potts, S.G.,Ramsay, A., Roberts, S.P.M., Woodcock, B.A.,Hopkins, A., Gundrey, A., Dunn, R., Tallowin, J.,Vickery, J., and Gough, S. (2006). Review of thediet and micro-habitat values for wildlife and theagronomic potential of selected plant species.English Nature Research Reports, No 697.Peterborough: English Nature.

Peach, W.J., Dodd, S., Westbury, D.B., Mortimer,S.R., Lewis, P., Brook, A.J., Harris, S.J., Kessock-Philip, R., Buckingham, D.L. and Chaney, K.(2011). Cereal-based wholecrop silages: Apotential conservation measure for farmland birdsin pastoral landscapes. Biol. Cons. 144: 836-850.

Policy Commission (2002) Farming and Food: ASustainable Future, Policy Commission on theFuture of Farming and Food, Cabinet Office,London.

Potts, S.G., Woodcock, B.A., Roberts, S.P.M.,Tscheulin, T., Ramsay, A.J., Pilgrim, E., Brown,V.K. and Tallowin, J.R. (2009). Enhancingpollinator biodiversity in intensive grasslands. J.Appl. Ecol. 46: 369-379

Woodcock, B.A., Edwards, A.R., Lawson, C.S.,Westbury, D.B., Brook, A.J., Harris, S.J., Brown,V.K. and Mortimer, S.R. (2008). Contrastingsuccess in the restoration of plant andphytophagous beetle assemblages of species-richmesotrophic grasslands. Oecologia 154: 773-783.

Woodcock, B.A., Potts, S.G, Tscheulin, T.,Pilgrim, E., Ramsey, A.J., Harrison-Cripps, J.,Brown, V.K. and Tallowin J.R. (2009). Responsesof invertebrate trophic level, feeding guild andbody size to the management of improvedgrassland field margins. J. Appl. Ecol. 46: 920-929

Woodcock, B.A., Vogiatzakis, I., Westbury, D.B.,Lawson, C.S., Edwards, A.E., Brook, A.J., Harris,S.J., Lock, K., Masters, G.J., Brown, V.K. andMortimer, S.R. (2010). The role of managementand landscape context in the restoration ofgrassland phytophagous beetles. J. Appl. Ecol. 47:366-376.

Improving floral and invertebrate diversity ingrassland field margins

R. Fritch1, 2, H. Sheridan1, J.A. Finn2, and D. ÓhUallacháin2

1UCD School of Agriculture, Belfield, Dublin 4,2Environment Research Centre, TeagascJohnstown Castle, Wexford.Email: dohuallachain@teagasc.ie

IntroductionPlant and invertebrate diversity within grasslandfarming systems has been declining, and protectionof biodiversity is one of the aims of the CAP. Fieldmargin management is a common measure in agri-environment schemes, but most research has beenconducted in arable conditions. Grassland fieldmargins are not as well researched as those inarable systems. Plant diversity in grassland fieldmargins is usually limited by the impoverishedseed bank, high soil nutrient status, and dominanceby rank vegetation (Pywell et al., 2002; Critchleyet al., 2002). This experiment investigates theestablishment and management methods toincrease plant and invertebrate diversity withingrassland field margins.

Materials and MethodsThe experimental site was located on the dairyfarm of the Teagasc research centre at JohnstownCastle, Co. Wexford. A stratified randomised split-plot field margin experiment was established inspring 2002. Nine 90 m long strips of grass swardalong existing fences were fenced off from thesurrounding paddock. One of three field marginwidths (1.5m, 2.5m and 3.5m) was randomlyassigned to each strip. Three field marginestablishment methods were randomly assigned tothree plots of 30m in length within a 90m strip.The establishment methods (Est. method) were: (1)fenced only; (2) rotavated and allowed toregenerate naturally; (3) rotavated and reseededwith a grass and herb seed mixture (n=41 species).Three replicates of each combination of width andestablishment treatments were made. Grazing wasintroduced to half the length of each 30m plot inJune 2003. Control plots consisted of existingpasture vegetation which had no subsequentapplication of nutrients or herbicide. Ungrazedplots were cut annually in September and theharvested biomass removed.

Sampling and analysisBotanical data were collected using permanentnested 1m x 3m quadrats in which presence ofplant species was recorded in July of 2003, 2007and 2008. Invertebrates were sampled in four ofthe treatments, using emergence traps, from Mayto August in 2007 and 2009. Spiders and parasitic

wasps were identified to species level. Speciesrichness and abundance data were analysed usingthe PROC GLIMMIX statement of SAS usingrepeated measures.

Results and DiscussionA total of 64 higher plant species were recorded,this included 43 herb, 15 grass, three tree, tworush, and one sedge species. Establishment methodand grazing had a significant effect on plantspecies richness (Table 1). Reseeded plots hadhighest species richness in all sampling periods (p< 0.0001). Grazing significantly increased quadratspecies richness (p< 0.0001). The interactionbetween these two factors over time was alsohighly significant (Table 1) as the grazed half ofplots increased in species richness over time whilethe ungrazed half of rotavated and reseeded plotsremained unchanged (Fig. 1)

Table 1. Effects of establishment, year, grazing,and width and the interactions of these on the plantspecies richness in 2003, 2007 and 2008.Significance levels: ns, not significant; *, p<0.05;**, p, 0.01; ***, p< 0.005.

Factor DF F Significance.Establishment (2, 68) 366.89 ***

Grazing (1, 76) 68.15 ***

Grazing xestablishment

(2, 76) 6.93 **

Width (2, 68) 1.23 n.s.

Width xestablishment

(4, 68) 2.83 *

Year x establishment (4, 68) 9.18 ***

Year x grazing (2, 76) 9.24 ***

Fig. 1. Mean plant species richness ( s.e.) withineach field margin establishment method over eachof three sampling periods (2003, 2007 and 2008).

(a)

(b)

Fig. 2. Mean plant species richness ( s.e.) withineach field margin establishment method over eachof three sampling periods (2003, 2007 and 2008).

Table 2. Effects of treatment on abundances ofdifferent invertebrate groups over six samplingperiods (2007 & 2008) in ungrazed marginscompared to controls.

Effect of establishment method

Abundance DF F Sig.

Spiders (3, 102) 42 ***

Wasps (3, 102) 15.9 ***

A total of 7,473 parasitic Hymenoptera individualswere trapped, comprising 132 genera from 16families. A total of 2,902 spiders were trapped.These included 43 species of mature spiders (n =816).

Fenced field margins (regardless of establishmenttreatment) resulted in increased abundance ofinvertebrate groups, in comparison to the grazedcontrol (Fig. 2, Table 2). This may be due to thereduction/exclusion of grazing pressure withinthese treatment plots, which facilitated thedevelopment of a more diverse sward architecture.These results concur with many arable field marginstudies which show that the provision of anuncropped field margin enhances invertebrateabundance when compared to a cropped margin

(Denys & Tscharntke, 2002; Thomas & Marshall,1999). In arable systems, leaving an uncroppedfield margin requires that the area adjacent to thefield boundary remains uncut, unploughed andunsprayed. However, as grasslands are generallygrazed by livestock, the development of anuncropped margin requires the installation offencing in order to exclude grazing, as well asexcluding inputs from livestock.

ConclusionsThe use of wildflower and grass seed mixtures wasmost successful in establishing a botanicallyspecies-rich field margin habitat. This would bemost appropriate in conditions where there is noexisting field margin flora, and the objective ishabitat creation. In this experiment, grazing had asignificantly positive effect on plant speciesrichness over time.

However, reduced grazing was required forincreased invertebrate abundances. No singleestablishment treatment was best for overallinvertebrate abundance as each taxa respondeddifferently.

Management methods for increasing one taxa mayconflict directly with the managementrequirements for other taxa. For example, mowingand grazing management increases plant species-richness in most grasslands; however, themaintenance of invertebrate diversity may requiretaller vegetation with less disturbance.

AcknowledgmentsFinancial support for this project was provided bythe Research Stimulus Fund of the Department ofAgriculture & Food.

ReferencesCritchley, C. N. R., Chambers, B. J., Fowbert, J.A., Sanderson, R. A., Bhogal, A. & Rose, S. C.(2002). Biological Conservation. 105:199-215

Denys, C. and Tscharntke, T. (2002) Plant-insectcommunities and predator-prey ratios in fieldmargin strips, adjacent crop fields, and fallows.Oecologia 130(2), 315-324.

Pywell, R. F., Bullock, J. M., Hopkins, A., Walker,K. J., Sparks, T. H., Burke, M. J. W. & Peel, S.(2002). Journal of Applied Ecology. 39:294-309

Thomas, C.F.G. and Marshall, E.J.P. (1999)Arthropod abundance and diversity in differentlyvegetated margins of arable fields. AgricultureEcosystems & Environment 72(2), 131-144.

a

c

bb

a

b

b

b

What is the conservation potential of grasslandson lowland farms?

C.A. Sullivan1, M. Sheehy Skeffington2 , J.A. Finn3

and M. J. Gormally4

1Humboldt State University, Arcata, CA, 95521,USA, 2 Botany and Plant Science, NUI Galway,3Teagasc, Environment Research Centre,Johnstown Castle, Co. Wexford, 4 Applied EcologyUnit, NUI Galway.Email: caroline.sullivan1@gmail.com

IntroductionSemi-natural grasslands and practices such ashaymaking were much more widespreadthroughout Ireland in the past (Stowe et al., 1993).Since 1990, semi-natural grassland, heathland andwetlands and mixed farmland have all decreased,while arable and permanent pastures increased byover 30% (EPA, 2006). Redressing the negativeimpacts of intensification on biodiversity has beenone of the major European agri-environment policydrivers (Henle et al., 2008). As a result, marginalgrasslands of moderate conservation value that arenot intensively managed have recently gainedrecognition, even though they may have beenreseeded in the past (Beaufoy, 2008). Though notexemplary species-rich meadows, they are oftenmore extensively farmed and provide importantecological services. The identification of thesegrasslands is necessary to accurately assess currentbiodiversity levels on farms. This has implicationsfor grassland conservation and restoration andwould be particularly useful for High Nature Value(HNV) farmland identification and monitoring, arequirement under the current Rural DevelopmentPlan (2007-2013).

Materials and MethodsTen percent of farms were selected randomly,outside nature designation sites, from six differentDistrict Electoral Divisions (DEDs) in east Co.Galway (a total of 603 fields). A W-shape waswalked through each field and the abundance ofeach plant species present recorded using theDAFOR (Dominant, Abundant, Frequent,Occasional, Rare) scale.All grassland types within the field boundarieswere identified according to Fossitt (2000).Nonmetric Multidimensional Scaling (NMS)analysis was carried. NMS was chosen as it avoidsthe assumption of linear relationships amongvariables and allows the use of distance measuressuited to non-normally distributed data (McCuneand Grace, 2002). Average Ellenberg values,indicators of plant species preference for positionsalong ecological gradients, were calculated forEllenberg N as a measure of each species’ responseto nitrogen (Hill 1999). This was done for each

field as a community calculation, using speciesEllenberg N values (1-9) and abundance. A one-way ANOVA was used to test for differencesbetween the mean species richness of the grasslandgroupings (log transformed to achieve normality).For further details see Sullivan et al. (2010).

Figure 1. NMS ordination A) with habitat typesaccording to Fossitt (2000) and B) with averageEllenberg N for each field.

Results and DiscussionAxis 1of the NMS ordination explained 33.9% ofthe variation in the matrix and Axis 2 explained38.9% (Fig 1). The majority of the fields (>70%)were classified as Improved Agricultural Grassland(sensu Fossitt, 2000) (Fig1A). However theordination indicated a continuum from low speciesrichness to high species richness from the bottom-left to the top-right of the ordination. Thiscontinuum was reinforced when the Ellenberg Nvalues were overlain on the ordination (Fig. 1B).An intermediate grassland habitat between semi-

natural and improved agricultural grassland (Fig.1)was also evident.

Grasslands with an average Ellenberg N value of 5are considered semi-improved (Robertson et al.,2002). On this basis, the fields were divided into 3groups: a semi-natural grassland group (EllenbergN < 4); a semi-improved grassland group(Ellenberg N = 5) and an improved agriculturalgrassland group (Ellenberg N ≥ 6). Species richness of the fields in each group differedsignificantly, F(2, 600) = 131.776, P < 0.0001.

Semi-improved grassland indicator plant species inBritain include Rumex acetosa, Ranunculus acris,Trifolium pratense, Cardamine pratensis, Prunellavulgaris, Achillea millefolium, Phleum pratenseand Anthoxanthum odoratum (DEFRA, 2005). Theoccurrence and abundance of many of theseindicators in this study showed that they would besuitable indicators of semi-improved grassland forIrish grasslands (Fig. 2). Rumex acetosa,Ranunculus acris and Trifolium pratense,Cardamine pratensis, Prunella vulgaris, Achilleamillefolium, Phleum pratense and Anthoxanthumodoratum consistently occurred in the semi-improved grassland group, but occurred in no morethan 10% of the improved grassland group.Plantago lanceolata, Cardamine pratensis,Prunella vulgaris, Leontodon autumnalis, Achilleamillefolium and Phleum pratense occurred morefrequently in the semi-improved grassland groupthan in any of the other groups.

Fig. 2. NMS ordination showing plant speciesdistributions. Black dots represent indicators ofsemi-improved grassland habitat (DEFRA, 2005).

The evidence supports the inclusion of a ‘Semi-Improved Grassland’ category in the Irishgrassland classification guidelines. Theidentification of semi-improved grassland areas inIreland has important conservation implications.Using this refined classification will more

accurately quantify grassland biodiversity. Formany lowland farms, using this classificationacknowledges the higher biodiversity value ofthese grasslands compared with more intensivelyfarmed fields. On some farms all fields previouslyidentified as improved agricultural grassland couldnow be re-classified as semi-improved grasslandfields. This reclassification is important whenconsidering both quantitative measures ofagricultural intensification and appropriatetargeting of agri-environment schemes. This isbecause it would provide an important resource forfuture restoration projects aimed at semi-naturalgrassland. It would also aid the identification ofHNV farmland, particularly at a stage when a farmhas already been identified as having HNVfarmland potential. At this stage the field-by-fieldapproach would provide the basis for identifying agrassland-based HNV farming system.

AcknowledgementsThis work was funded by Teagasc under the WalshFellowship Scheme.

ReferencesBeaufoy, G. (2008) HNV Farming -Explaining theconcept and interpreting EU and national policycommitments. EFNCP, UK.

DEFRA, (2005) Environmental Stewardship FarmEnvironment Plan Guidance 008. Identification ofGrassland Features. Department for Environment,Food and Rural Affairs, UK.

EPA, (2006) Environment in Focus 2006:Environmental Indicators for Ireland.Environmental Protection Agency, Wexford.

Fossitt, J. (2000) A Guide to Habitats in Ireland.The Heritage Council, Kilkenny.

Henle et al. (2008) Identifying and managing theconflicts between agriculture and biodiversityconservation in Europe. Agric., Ecosys. & Env.124: 60-71.

McCune, B., Grace, J.B., 2002. Analysis ofecological communities. MjM Software Design.

Robertson et al. (2002) Monitoring grasslandbiodiversity action plan targets. Final Report.English Nature.

Sullivan, C.A., Skeffington, M.S., Gormally, M.J.,Finn, J.A., 2010. The ecological status ofgrasslands on lowland farmlands in western Irelandand implications for grassland classification andnature value assessment. Biol. Cons. 143: 1529-1539.

Stowe, T.J., Newton, A.V., Green, R.E., Mayes, E.,(1993) The decline of the Corncrake Crex crex inBritain and Ireland in relation to habitat. J. App.Ecol. 30: 53-62.

High nature value meadows: results from anational grassland survey

J.R Martin, F.H. O’NeillBotanical Environmental & ConservationConsultants, 26 Upper Fitzwilliam St., Dublin 2Email: jmartin@botanicalenvironmental.com

IntroductionA comprehensive study of semi-natural grasslandhabitats is currently being carried out in Ireland(O’Neill et al., 2009). In addition to classifying allsurveyed grasslands within a Fossitt (2000) habitatcategory, the study is developing a classificationfor Irish semi-natural grasslands. Meadow systemscomprise a significant part of this study, and manycan be considered to be of high nature value(HNV) by virtue of their species diversity andstructure. A further aim of the study is to assessgrassland habitats that correspond to an EUHabitats Directive Annex I habitat; these may beconsidered as a particularly diverse sub-set ofHNV grasslands with specific characteristics interms of their species composition and structure.The two Annex I meadow habitats recorded inIreland are the wet meadow habitat Moliniameadows (6410) and the dry meadow habitatLowland hay meadows (6510). Under the EUDirective, Ireland has a responsibility to protectthese habitats and maintain them at a favourableconservation status.

Materials and MethodsField methodologyA 2 m x 2 m relevé was recorded from within eachgrassland habitat located within a site. Multiplerelevés were recorded where there was significantvariation in the floral composition within a habitattype. For each relevé, the grid reference,topography, altitude, slope and aspect wererecorded and a soil sample was collected foranalysis of pH, organic content and totalphosphorus. Sward characteristics such as forb tograminoid ratio and sward height were also noted.Cover in vertical projection for each vascular andbryophyte species was recorded on the Dominscale, as were other general parameters such asbare soil, leaf litter, total field layer and totalbryophyte cover. For full details of themethodologies used see O’Neill et al. (2009).

Vegetation analysisOutlier analysis was carried out on the 2007-2010relevé dataset using PC-Ord 5. Relevés outside theremit of the study were omitted and all remainingrelevés were analysed using the twocomplementary statistical techniques ofHierarchical Polythetic Agglomerative Cluster

Analysis and Indicator Species Analysis (O’Neillet al., 2009).Management dataFrom 2010 the recently devised methods forscoring impacts and activities within an Annex Ihabitat was utilised (Ssymank, 2010).

Results and DiscussionSince 2007, 3,078 relevés have been recorded in783 sites covering 17,814 ha of semi-naturalgrassland, much of which is of HNV. The 783 sitesare spread over 12 counties and all four provincesof Ireland. A total of 3,024 relevés were analysedto produce a classification of Irish grasslands thatrecognises 34 grassland vegetation types. Theoccurrence of the two Annex I meadowcommunities (6410 and 6510) within the 34vegetation types was examined. Only those typeswhich had a significant number of relevés classedas Annex I meadows were identified as beingimportant for HNV meadows. In all, 75% (53 of 71relevés) of relevés of the rarer dry meadow AnnexI habitat 6510 were found to occur across four drygrassland types, while 79% (242 of 306 relevés) ofrelevés of the wet Annex I habitat 6410 werepresent in seven wet grassland types. These fourdry vegetation types (represented by 189 relevés)were therefore taken to include the majority ofHNV dry meadow communities and the seven wettypes (represented by 735 relevés) to include themajority of HNV wet meadow communities.The environmental differences between the twomeadow types can be largely attributed to edaphicfactors. Dry meadows were recorded primarily onwell-drained mineral soils (58%) and gleys (26%).Wet meadows were recorded on gleys (61%) andpeats (26%). Soils and other parameters aredefined for the two types of meadow in Table 1.

Table 1. General relevé data for wet and dry HNVmeadows. The main two soil types are shown(WDM=Well-drained mineral) and the meanspecies richness, grass height (cm), forb height(cm), forb proportion (%) and slope (degrees).Data type Dry

meadowsWetmeadows

Soils 58% WDM,26% Gleys

61% Gleys,26% Peats

Sp. richness 23 23Grass height 29 cm 41 cmForb height 24 cm 25 cmForb prop. 47% 31%Slope 4º 3º

As the data presented show, apart from soil themost pronounced differences between wet and drymeadows are the increased proportion of forbs indry meadow communities, due to a high incidenceof species such as ribwort plantain (Plantagolanceolata) and yellow rattle (Rhinanthus minor),

and the generally taller grass sward in wetmeadows, often due to the presence of purplemoor-grass (Molinia caerulea).The differences between the Annex I meadow(6410 and 6510) communities and the non-AnnexHNV meadows are subtle, with the Annex Imeadows tending to have taller forb height and ahigher proportion of forbs (Table 2). The speciesrichness of Molinia meadows (6410) is noticeablyhigher than the non-Annex HNV wet meadows.

Table 2. Comparison of wet and dry non-AnnexHNV meadows with their corresponding Annex Ihabitat relevé data. Shown are mean speciesrichness, grass height (cm), forb height (cm), forbproportion (%), and slope (degrees)Data type Dry

meadowsWetmeadows

HNV 6510 HNV 6410Sp. richness 23 23 22 25Grass height 29cm 29cm 40cm 43cmForb height 22cm 28cm 24cm 28cmForb prop. 43% 58% 29% 37%Slope 5º 2º 3º 3º

Although it was expected that mowing would bethe most common management type for HNVmeadows, this was not always the case. ForAnnex I dry meadows (6510) they were alwaysmown, or mown in combination with grazing.However, Annex I wet meadows (6410) were onlymown, or mown in combination with grazing in26% of cases (Table 3). Mowing is more prevalentin all sites than Molinia meadows (6410), whichdoes raise issues regarding the future of thisimportant habitat. The fact that 19% of Moliniameadows are abandoned is also of some concern asmeadows require management to maintain them.

Table 3. Management data for all 2010 grasslandsites (n=203), Annex I habitats 6410 (n=31) and6510 (n=6).Management All

sites6510 6410

Mowing only 5% 50% 7%Grazing only 59% 0% 55%Mowing &grazing

29% 50% 19%

None recorded 7% 0% 19%

A decline in the cutting of HNV wet meadows hasprobably occurred over the last 50 years due toincreased mechanisation and the difficulty ofcutting in wet environments with heavy machinery.

It is generally accepted that many former HNVmeadows have been improved over the last 50years through a combination of drainage, fertiliserapplication, and the planting of high yieldingspecies such as Lolium perenne. The more

productive systems that have replaced HNVmeadows have a lower species diversity and neverequate to the ecologically important Annex I wetand dry meadows. Although HNV meadows areless productive, there is anecdotal evidence that thespecies-rich hay cut from them fetches a higherprice, which would off-set some losses in yield.However, the lower yields of HNV and Annex Imeadow systems, compared to improvedagricultural swards, necessitates some grantscheme or other incentives for farmers if theseimportant meadow systems are to be maintained.

ConclusionsOur data have shown that there is a diversity ofHNV meadow communities in Ireland, with fourtypes of dry meadow and seven types of wetmeadow identified. Annex I meadow habitats arerare in Ireland, with only 12% of all relevésrecorded between 2007 and 2010 classed asAnnex I Lowland hay meadows (6510) or Moliniameadows (6410). Dry HNV meadows are evenrarer than wet HNV meadows, probably due to theease with which these meadow systems can beimproved. Lack of appropriate management raisesconcerns about the future prospects of manyMolinia meadows. Of particular concern is achange in the structure and functions of the habitatwhich will have negative implications for plantspecies diversity and fauna, such as the EU AnnexII species marsh fritillary.

AcknowledgmentsThis research is funded by the National Parks andWildlife Service, part of the Department of theEnvironment, Heritage, and Local Government.We would also like to thank Marie Dromey ofNPWS for providing project direction and ensuringthe continued funding for the study. Thank you tothe botanists who have worked on the project andin particular Philip Perrin, Fiona Devaney, KateMcNutt and Aoife Delaney who helped collect thedata and contributed to the annual report that thispaper is based on.

ReferencesFossitt, J.A. (2000). A Guide to Habitats inIreland. The Heritage Council, Kilkenny.O’Neill, F.H., Martin, J.R., Perrin, P.M., Delaney,A., McNutt, K.E. & Devaney, F.M. (2009). IrishSemi-natural Grasslands Survey Annual ReportNo.2: Counties Cavan, Leitrim, Longford andMonaghan. Report submitted to NPWS, Dublin.Ssymank, A. (2010). Reference list Threats,Pressures and Activities (final version).http://circa.europa.eu/Public/irc/env/monnat/library?l=/expert_reporting/work-package_revision/sub-group_papers/pressures_threats&vm=detailed&sb=Title. Accessed 9/3/2011.

Identification and distribution of HNVfarmland in Germany

D. Fuchs1 and A. Benzler2

1PAN Planungsbüro für angewandten NaturschutzGmbH, Rosenkavalierplatz 10, 81925 München,Germany, 2Federal Agency for NatureConservation (BfN), Konstantinstraße 110, 53179Bonn, GermanyEmail: Daniel.Fuchs@pan-gmbh.com

IntroductionIn early cartographical estimations of the potentialamount of High Nature Value (HNV) farmlandwithin the EU (e.g. EEA 2004), Germany (as wellas Northern France, Denmark, and the LowCountries) showed almost no such farmland at all.The data sources available on the EU level forthese estimations (e.g. CORINE land cover data)clearly failed to discern potential HNV farmland inthese regions, where intensely used agriculturalareas are interspersed with relatively smallremnants of species-rich, often traditionally usedplots. In order to correctly estimate the amount ofHNV farmland in Germany, a working groupcomprising of federal and regional conservationand agricultural authorities and three privateenvironmental consultancies, decided in 2008 ondeveloping a nation-wide sampling programme.This programme was to yield the data necessary toreport on the HNV farmland indicator (baselineindicator 18) under the current EU CommonMonitoring and Evaluation Framework (CMEF). Afirst round of surveying took place in 2009 withrepeat surveys having taken place in 2010 andbeing planned for 2011 to 2013.

Materials and MethodsThe German HNV farmland survey uses a samplebased approach which is structurally similar to theBritish Countryside Survey (Carey et al., 2008).Fieldwork, to locate HNV farmland structures,took place on square sample plots of 1 km² area.Field workers were furnished with a manual,comprising of keys to identify different types ofHNV farmland and to evaluate their “nature value”on a three-tier scale.

Sampling designWe used a sample design which had beendeveloped by the Federal Statistical Office inconjunction with the Federal Agency for NatureConservation for conservation surveys in generaland has been used for several years for the GermanCommon Birds Census (GCBC) (Mitschke et al.,2005). The sample consists of 1,000 squares of1km² size and is stratified using 21 so-calledecoregions, i.e. areas defined by a combination ofsoil, vegetation, climate and elevation

characteristics (Schröder and Schmidt, 2001), andsix main land use categories to define the strata.Financial constraints forced the removal of 118squares with more than 95 % non-agricultural use(mainly woodland and settlements) from thesample. On the other hand, two Länder decided toincrease their samples and use them for otherpurposes than HNV farmland monitoring as well,resulting in an overall sample of 916 squares forthe field season 2009.

Field surveysField work in 2009 was commissioned toexperienced ecological surveyors in lotscomprising 10 – 15 squares each. Most of the fieldwork took place in May to July 2009 withadditional field work up to September 2009. Thefield manual defines 16 types of HNV farmlandincluding: species rich arable, grassland, orchards,vineyards, and fallow land on the one hand, and 10types of landscape elements such as hedgerows,copses, dykes, reedbeds, sunken lanes, dry stonewalls one the other hand. In addition, all habitattypes listed on Annex I of the EU Habitatsdirective are counted as HNV farmland elements,if they are dependent on at least occasionalagricultural use (e.g. type 5130 “Juniperuscommunis formations on heath”).Two sets of keys were used to identify HNVfarmland in the field. For arable, grassland,orchards, vineyards, and fallow land, lists ofindicator taxa were used, comprising 12(vineyards) to 38 (grassland) taxa each. Taxa wereeither species (e.g. Great Burnet Sanguisorbaofficinalis for grassland), genera (e.g. PoppyPapaver sp. for arable) or morphospecies (e.g.“small yellow-flowered clover” for grassland). Anarea was identified as HNV farmland when at leastfour of the above indicator taxa were found within1 m on both sides of a 30 m transect in the area.For landscape elements, a qualitative key wasdeveloped using mainly structural criteria such aswidth and number of woody plant species forhedgerows.All HNV farmland elements found in the fieldwere sorted in three categories of “moderately highnature value”, “very high nature value” and“extraordinarily high nature value”. Thisassessment used the criteria mentioned above withe.g. the levels for grassland being defined by 4-5,6-7 and 8 and more indicator taxa within the 30 mtransect.Field data were digitized as polygons using arealphotographs of 1:5,000 scale for boundarydefinition. HNV type and nature value were storedin a database for each polygon, as were all data onindicator taxa found on each transect.

Results and DiscussionIn all squares, 43,370 HNV farmland parcels andelements were found, comprising an area of13,144.65 hectares. Species rich grassland was thedominant HNV farmland type with 4,328.26hectares (32.9% of the total), followed byhedgerows (955.90 hectares or 7.3%) and species-rich arable (757.99 hectares or 5.8%).These raw numbers, however, tell relatively littleabout the proportion of HNV farmland on allagricultural area. An estimation of the proportionof HNV farmland on agricultural area (AA) has toconsider the sample stratification, including thefact that some Länder increased the sample size.Taking these factors into account, an estimate of25,103.90 km² HNV farmland for the whole ofGermany was calculated, which translates into13.0% (with an absolute sampling error of ± 0.4%)of agricultural area. Nature value of HNV farmlandis spread unequally with 6.3% ± 0.3% of AA beingof “moderately high nature value”, 4.5% ± 0.2% of“very high nature value” and the remainder of2.1% ± 0.1% of “extraordinarily high naturevalue”. The estimates for the different HNVfarmland types are collated in Table 1.

Table 1. Proportion of different types of HNVfarmland on agricultural area of Germany.Abbreviations: Prop. AA = proportion onagricultural area, a.s.e = absolute sampling error.HNV farmland type Prop. AA a.s.e.Arable 1.5% 0.1%Grassland 5.7% 0.3%Fallow 0.8% 0.1%Orchards 0.7% 0.1%Landscape elements 4.3% 0.1%

Most of the HNV landscape elements amount toless than 0.5% of AA, the exceptions beinghedgerows with 1.1% ± 0.04% and dykes with0.5% ± 0.03%. Vineyards could not be adequatelysampled in 2009 since field work started only inMay, when too many of the geophytes, which aretypical for traditionally used and species-richvineyards, were past flowering.The estimated amount of HNV farmland differsstrongly between the ecoregions used in samplestratification. Highest proportions (20.4 to 23.7%)were found in the higher parts of mountains insouthern and central Germany (e.g. Black Forest,Bavarian/Bohemian Forest, Harz) and in the lakesregion in north-eastern Germany. Ecoregions withthe lowest proportions of 7.0 to 8.4% are situatedin the heavily populated Ruhr and Rhein/MainValleys and in parts of Germany with rich soilsand long traditions of large-scale agriculture.On the level of the ecoregions, the proportion oftotal HNV farmland is strongly correlated with theproportion of HNV grassland (Spearman’s rank

correlation test, ρ = 0.83, p<0.001) which in turn is negatively correlated with the amount ofagricultural area on total area of the ecoregion (ρ = –0.53, p<0.05). On the other hand grassland showsno significant correlations with any of the otherHNV farmland types, either singly or groupedtogether.Generally, these results confirm some keyassumptions about the situation of HNV farmlandin Germany. Grassland seems to be the mostimportant factor, since more than 43 % of all HNVfarmland is grassland and since grassland tends tobe more species rich in areas where there (still) is alot of it. Also, mountainous regions, with relativelyhigh woodland amounts and a low proportion ofagricultural area, tend to have the highest HNVfarmland shares.

ConclusionsThe sampling programme outlined here allows anestimation of the amount of HNV farmland inregions were such farmland is predominantlyfound in intensely used areas. These HNVfarmland elements contribute decisively to localbiodiversity. In the future, the monitoring of thechange in their distribution and total amount underincreasing pressure from agriculturalintensification will be the most important task ofthe monitoring programme.

AcknowledgmentsWe wish to thank all representatives of the Länderconservation and agriculture authorities who in theintroduction of the monitoring programme andespecially Rainer Oppermann (IFAB) and AlfonsKrismann (ILN) who developed crucial aspects ofthe methodology with us.

ReferencesCarey, P.D., Wallis, S., Chamberlain, P.M.,Cooper, A., Emmett, B.A., Maskell, L.C., et al.(2008). Countryside survey: UK results from 2007.NERC/Centre for Ecology & Hydrology.

EEA European Environment Agency (2004). Highnature value farmland : characteristics, trends,and policy challenges. EEA reports. EuropeanEnvironment Agency, Copenhagen.

Mitschke, A., Sudfeldt, C., Heidrich-Riske, H., andDröschmeister, R. (2005). Das neue Brutvogel-monitoring in der Normallandschaft Deutschlands–Untersuchungsgebiete, Erfassungsmethode underste Ergebnisse. Vogelwelt 126: 127–140.

Schröder, W. and Schmidt, G. (2001). Definingecoregions as framework for the assessment ofecological monitoring networks in Germany bymeans of GIS and classification and regressiontrees (CART). Gate to Environmental and HealthScience 3: 1–9.

Habitats in the Irish farmed landscape

H. Sheridan1, B. Keogh1, A. Anderson1, T.Carnus1, B.J. McMahon1, S. Green2 and G. Purvis1

1 School of Agriculture, Food Science andVeterinary Medicine, University College Dublin,Belfield, Dublin 4, 2 Teagasc, Kinsealy ResearchCentre, Dublin 17.Email: Helen.sheridan@ucd.ie

IntroductionIt is now widely accepted that the continuedexistence of many species, including those whichare still relatively common, is heavily dependentboth on the maintenance of diverse agriculturalpractices, and the retention of a matrix of semi-natural habitat within the farmed landscape(Donald and Evans, 2006). Agriculturalcommunities are coming under increasing pressureto justify the continuation of direct payments suchas the Single Payment Scheme, through theprovision of public goods. Farmland habitats andthe biodiversity which they support represent onesuch public good.However, surprisingly little attention has beenafforded to the classification and quantification ofhabitats at farm scale in Ireland. This study seeksto address this deficiency through the provision ofa baseline dataset of farmland habitats againstwhich the relationships between geographicallocation, farming practice and AE Schemeparticipation can be assessed. Provision of thisbaseline data also creates the potential to allowevaluation of the likely influence of futurestructural changes in Irish agriculture andproduction systems, ongoing development of agri-environmental measures and potentially the longer-term effects of global warming on Irish farmlandhabitats.

Materials and MethodsHabitat surveys were undertaken on 118predominately pastoral farms. Farms were locatedin three regions i.e. Sligo-Leitrim (n = 39), Offaly-Laois (n = 40), Cork (n = 39). This constitutes anorth-south geo-climatic gradient across thecountry, with increased farm management intensityand a converse decrease in REPS participation.Five 10km2 were randomly selected from eachregion during 2007 and 2008. The four central 1kmsquares within each 10km square were identifiedand a farm, located as close as possible to themiddle of each of these central squares wasrandomly selected.

Sampled farms were classified by REPSparticipation status, farm system i.e. Dairy, Beef,Suckler and ‘Other’ (which included a miscellanyof essentially extensive management types), andstocking rate (LU ha-1).

The type and extent of all habitats was recordedonto farm maps. Classification of habitatsprincipally followed the designations of Fossitt(2000). However, grassland fields on each farmholding were walked and a comprehensive, thoughnot exhaustive visual assessment of componentplant species was recorded according to theDAFOR scale. Fields were subsequently assignedto the following classification of agriculturalgrassland types:

a) Intensively managed grassland (GA1): Loliumperenne was ≥ 70% of sward cover.

b) Improved grassland (Grade 1) (GA1): L.perenne may have been dominant but aminimum of four other species were also foundto occur frequently within the sward.

c) Improved grassland (Grade 2) (GA1): L.perenne only of occasional/rare occurrence, ifpresent within the sward. These grasslandstypically contained ten to fifteen frequentlyoccurring species.

d) Transitional grassland-scrub (not recognised):Significant incidence of woody species suchas: Prunus spinosa and Ulex europaeus.

e) Species rich wet grasslands (GS4): Anabundance of frequently occurring herbs e.g.Cirsium palustre, Lychnis flos-cuculi, Angelicasylvestris and Filipendula ulmaria etc.

f) Juncus dominated wet grasslands (GS4).g) Wet grassland (GS4)- Juncus spp frequent but

not dominant with a number of grass and herbspecies. Generally evidence of improvement

h) Callows (seasonally inundated) (GS4)

Farm habitats were digitised onto OSIorthophotographs (2004) and their extentquantified using ArcGIS software. Proportion ofthe surveyed farm area under different habitattypes were computed and categorised as follow:

a) Agriculturally productiveb) Agriculturally marginally productivec) Non-cropped semi-natural habitatd) Other

Data analysis

Generalized Linear Model analysis (GLMs) wasundertaken to investigate the influence of Region,Farm System and REPS participation status on thearea classified under each of the broad habitatcategories. Principal Components Analysis (PCA)was used to investigate the relationship betweenthe environmental variables and the proportion offarm area under individual habitat types.

Results and DiscussionCollectively, the 118 farms accounted for 5,673 hawith an average farm size of 48 ha (± 3.03 s.e.).Habitat surveys were undertaken on 3,688 ha with

an average area of 31 ha (± 1.88 s.e.) surveyed perfarm.A breakdown of the proportion of farm areaclassified as a) Agriculturally productive, b)Agriculturally marginally productive, c) Semi-natural and d) Other, across the three studyregions, four farming systems according to theirREPS participation status, is presented in Fig. 1.This indicates that greatest difference in farmhabitat structure was primarily due to the differingratios of agriculturally productive land tomarginally productive land recorded within each ofthe three study regions, with a greater proportionof marginally productive land recorded in Sligocompared with Cork (Fig. 1). However, thisregional effect also incorporates the farm systemeffects, as farm system was found to show a highregional dependency (Fig. 2). No relationship wasfound between farm habitat composition and REPSparticipation status.

Fig. 1. Habitat type as a proportion of surveyedfarm area across region, farm system and REPSparticipation.

The PCA biplot of habitats and environmentalvariables is presented in Fig. 2. This shows a closeassociation between dairy, non-REPS, increasingstocking rates, and the Cork region. Intensivelymanaged grassland was the only habitat typewhose incidence showed close association withthese variables. Most of the natural / semi-naturalhabitat types were closely associated with eachother and with the Sligo region and sucklerfarming systems, with REPS participation andduration also ordinating in this direction.Additional farm system types such as ‘other’,‘beef’ and ‘sheep’ ordinate separately from bothdairying and suckler systems, indicating that theirhabitat composition tends to be different. Ingeneral these appear to represent a ‘middle ground’in terms of farm management intensity. Grade 2

grassland was found to ordinate in their generaldirection.

Fig. 2. PCA biplots of farm habitat compositionwith passive projection of management variables,nominal variables are represented as centroids.Habitat types: JD grass, Juncus Dominated Wet Grassland; Wet gras,Wet Grassland; Imp G2, Improved Grassland Grade 2; Sprich, SpeciesRich Wet Grassland; FB, Field Boundaries; Trans, Transitionalgrassland to scrub; Till, Tillage/Arable; Int Gr, Intensive Grassland;Odecwood, Old Deciduous Woodland; Imp G1, Improved GrasslandGrade 1; SR, Stocking Rate; st_N, Standardised N farm input level;st_P, Standardised P farm input level.

ConclusionsThese results indicate that the diversity anddistribution of farmland habitats is largelydependent on region and farming system. Despitethe inseparable nature of these variables, ourresults indicate the overall need for targeting andcustomisation during the development of future AEpolicy in order to ensure maximum ecologicaleffectiveness. Our data also represent a hugelyimportant resource in the justification forcontinued SPS payments, and indicate that a muchhigher proportion of ecologically important habitattypes have been retained on Irish farmland whencompared with many other EU countries (seeManhoudt and de Snoo, 2003). However, theseresults also indicate the inadequacies of the currenthabitat classification guide (Fossitt, 2000) anddemonstrate the need to address theseinadequacies.

AcknowledgmentsThe authors thank DAFF for the provision offunding under the Research Stimulus Fund 2006.

ReferencesDonald, P.F. and Evans, A.D. (2006). Habitatconnectivity and the matrix restoration: the widerimplications of agri-environmental schemes. J.App. Ecol. 43: 209-218.

Fossitt, J.A. (2000). A Guide to Habitats inIreland. Heritage Council, Kilkenny.

Manhoudt, A.G.E., and de Snoo, G.R. (2003). Aquantitative survey of semi-natural habitats onDutch arable farms. Agric. Ecosyst. and Environ.97: 235-240.

Pollinators and pollination networks in Irishfarmland: implications for conservation ofpollination services

J.C. Stout, E.F. Power, D.A. Stanley and S.E.MullenBotany Department, School of Natural Sciences,Trinity College Dublin, Dublin 2, ROIEmail: stoutj@tcd.ie

IntroductionBees and other flower-visiting insects play animportant functional role as pollinators for bothcrops and wild plants worldwide. Hence theyprovide a key ecosystem service. However,declines in both the abundance and speciesrichness of pollinating insects have been recordedin Europe, North America and Asia (e.g. Williamsand Osborne, 2009). These declines are thought tobe primarily due to habitat loss, fragmentation anddegradation (Potts et al., 2010), often due toagricultural intensification. Changes in agriculturalpractices, including increased mechanisation,removal of semi-natural features such ashedgerows, simplification of landscapes,widespread use of agrichemicals, increasedstocking densities, reliance on a low number ofcommercial plant strains, frequent reseeding andincreased silage cutting, have resulted in loss offorage, nesting and mating sites for pollinatinginsects (Potts et al., 2010). In Ireland, some speciesof bumblebee (Bombus spp.) have declined andshifted westward to the extremity of their ranges(Fitzpatrick et al., 2007), most probably because ofchanges in agricultural practices. However, theeffects of land use change on other pollinatingtaxa, and on the interactions between flowervisitors and the plants they pollinate, are not wellunderstood. In particular, the effects of currentconventional farming practices compared withalternatives such as organic farming andcultivation of bioenergy crops, are not wellunderstood (Dauber et al., 2010; Power and Stout,2011). In this paper, we compare data from surveysof flower-visiting insects and plants from Irishfarmland and semi-natural grasslands across thecountry, in order to assess pollinator status and tomake predictions about potential impacts onpollination services.

Materials and MethodsSurveys of plants and pollinators and theirinteractions were made in a total of 65 sites acrosscentral, southern and eastern Ireland: 25 sites werepastures (10 organic, 15 conventional), 20 weresemi-natural grasslands, 10 were tillage crops(winter wheat and oil seed rape), and 10 wereenergy crops (Miscanthus). Surveys wereconducted during a two year period (2009-2010),

although each site was surveyed in one year only.Flower-visiting insects were sampled usingtransect walking methods. Flowering plant surveyswere conducted using transects and quadratmethods.

Interaction networksData matrices of the total number of interactionsbetween plants and flower visiting insects observedin each site were constructed. Bipartite visitationgraphs and network descriptors were calculated foreach site using the “networklevel” command in thebipartite package (Dormann et al., 2009) in R(version 2.11.1, R-Development-Core-Team2007).

Data analysisData were standardised for sampling effort (perunit time for insect observations and per unit areafor floral observations) and compared amongfarming types. Relationships between insect andplant variables were tested using Pearson’s ProductMoment Correlation Coefficient.

Results and DiscussionOf the 6,723 insects recorded visiting flowers inthe sites, 57% were hoverflies (Diptera,Syrphidae), 30% bumblebees (Hymenoptera,Apidae), 4% other bees (Hymenoptera, Apidae)and 8% butterflies (Lepidoptera). Flower visitinginsects were more abundant and species rich inMiscanthus (M), semi-natural grasslands (SN) andoilseed rape (OSR) sites and least abundant andleast species rich in wheat crops (W) (Figure 1aand b). Flower species richness was highest in SN(Figure 1c). Insect and flower species richnesswere significantly positively correlated with oneanother across all sites (p<0.01).Insect-flower interaction network parameters didnot vary greatly among site types (Table 1).However, low animal-plant ratios (which aretypically approximately 4:1 in other networks)suggest that there is little insect redundancy innetworks which are therefore less likely to betolerant to extinction (Memmott et al., 2004). Lowlevels of connectance are typical in plant-pollinatorinteraction networks, but less well-connectednetworks are less robust to species loss. Interactionstrength asymmetry was highest in SN andpastures (P). This parameter measures theimbalance between interaction strengths of aspecies pair (Dormann et al., 2009), with specialistflower visitors interacting with generalist plants(Bascompte et al., 2006). Networks in all siteswere quite highly nested (this parameter ismeasured as departure from perfectly nested matrix= 0), which suggests that less connected speciesinteract with a core of highly connected species.

0

5

10

15

20

25

Mea

npl

ant

spec

ies

rich

ness

0

50

100

150

Mea

nin

sect

abun

dan

cepe

rhou

r

0

5

10

15

Mea

nin

sect

spec

ies

rich

ness

per

hour

a

b

c

Figure 1. Mean (± SE) insect abundance (a), insectspecies richness (b), and plant species richness (c)per site type.

Table 1. Mean network parameters (AP = animalplant ratio, QC = quantitative connectance, ISA =interaction strength asymmetry, N = nestedness)per site type (M = Miscanthus, OSR = oilseed rape,P = pasture, PO = organic pasture, W = wheat,SN=semi-natural grassland).Site AP QC ISA N

M 1.14 0.14 0.03 21.5OSR 1.29 0.15 0.06 22.9P 1.51 0.19 0.13 21.2PO 1.32 0.19 0.10 22.5W 1.82 0.20 0.04 26.1SN 1.44 0.11 0.14 20. 5

ConclusionsWheat crops appear to provide the least resourcesfor pollinating insects, whilst semi-natural sitescontain the greatest floral diversity and higherinsect diversity and numbers. Miscanthus cropsappear to be relatively rich in pollinating insects,possibly because of their perennial, low-input

nature. However, low insect redundancy androbustness in all networks suggests that insect-flower interaction networks in all sites arevulnerable to insect species extinction, i.e. if insectspecies are lost from sites, this could have knock-on impacts on the plants they pollinate.Conservation measures to increase both floralresources and nesting opportunities would benefitpollinating insects, and improve communitystability. For example, sensitive hedgerowmanagement; reducing/eliminating herbicideusage, especially on uncultivated areas; sowingclover into pastures (and allowing plants toflower); cultivating wild-flower areas; and cuttingfor hay/silage later in the year after flowering hasceased. Whether these measures will improvepollination services remains to be resolved.

AcknowledgmentsThis work was funded by DAFF RSF (07/512),EPA SIMBIOSYS (2007-B-CD-1-S1) andIRCSET PG studentship to SEM. We are gratefulto land owners for allowing us to sample.

References

Bascompte, J., Jordano, P., Olesen, J.M. (2006)Asymmetric coevolutionary networks facilitatebiodiversity maintenance. Science 312, 431-433.

Dauber, J., Jones, M.B., Stout, J.C. (2010) Theimpact of biomass crop cultivation on temperatebiodiversity. Global Change Biol.: Bioenergy 2,289-309.

Dormann, C.F., Fründ, J., Blüthgen, N., Gruber,B.I. (2009) Indices, graphs and null models:analyzing bipartite ecological networks. OpenEcology J. 2, 7-24.

Fitzpatrick, U., Murray, T.E., Paxton, R.J., Breen,J., Cotton, D., Santorum, V., Brown, M.J.F. (2007)Rarity and decline in bumblebees - A test of causesand correlates in the Irish fauna. Biol. Cons. 136,185-194.

Memmott, J., Waser, N.M., Price, M.V. (2004)Tolerance of pollination networks to speciesextinctions. Proc. Roy. Soc. B – Biol. Sci. 271,2605-2611.

Potts, S.G., Biesmeijer, J.C., Kremen, C.,Neumann, P., Schweiger, O., Kunin, W.E. (2010)Global pollinator declines: trends, impacts anddrivers. TREE 25, 345-353.

Power, E.F., Stout, J.C. (in press). Organic dairyfarming: impacts on insect–flower interactionnetworks and pollination. J. Appl. Ecol. DOI:10.1111/j.1365-2664.2010.01949.x

Williams, P.H., Osborne, J.L. (2009) Bumblebeevulnerability and conservation world-wide.Apidologie 40, 367-387.

Monitoring of the Agri-environmentProgramme within the Northern Ireland RuralDevelopment Programme 2007-2013

M. Flexen, D. O’Mahony, and J.H. McAdamAgri-Food and Biosciences Institute (AFBI),Newforge Lane, Belfast, BT5 9PXEmail: mel.flexen@afbini.gov.uk

BackgroundSince the introduction of agri-environment (AE)schemes in Northern Ireland, the Department ofAgriculture and Rural Development (DARD) hasbeen obliged to monitor scheme performance inrelation to environmental objectives. Long-termscientific monitoring of AE schemes has beencarried out since 1993 (e.g. McEvoy et al., 2006)Previous schemes have now been replaced by theNorthern Ireland Countryside ManagementScheme (NICMS) which opened in 2008 (DARD,2008). The NICMS, together with the OrganicFarming Scheme (OFS), form the AE programmewithin the Northern Ireland Rural DevelopmentProgramme (NIRDP) 2007-2013.

All EU member states are required to monitorRDPs using the Common Evaluation andMonitoring Framework (CMEF). The CMEFspecifies the use of common indicators that mustbe included and monitored. Member states mayalso establish a limited number of programmespecific additional indicators. These shouldprovide information that both identifies theprogramme efficacy and informs fine tuning anddevelopment of measures for future AEprogrammes. The additional indicators andassociated targets for the AE programme inNorthern Ireland were developed by DARD inconjunction with stakeholders. A 3-year researchand monitoring project based on these indicatorscommenced in 2010. This paper briefly discussesaspects of current monitoring relating tobiodiversity on farmland under AE schemeagreement.

Biodiversity indicatorsFarmland birdsUsing existing baseline data, farm-scale surveyswill measure changes in the abundance of lapwingsand seed-eating farmland birds (includingyellowhammer) over a 5 year period on AE farmsand non-AE farm controls. Field-scale evaluationsof the usage of AE options, such as conservationcereal and wild-bird cover, are also being carriedout to assess the benefits for seed-eating birds insummer and winter. The survey work commencedin November 2010 and will continue until March2012.

Irish hareThis study will provide information on how the‘delayed cutting and grazing option’ within theNICMS may influence leveret survival. Thisoption provides farmers the opportunity to becompensated for delays in the cutting of silage orgrazing of livestock in pastures. This shouldbenefit Irish hares in terms of reduced disturbanceduring the peak of the breeding season (April-June). The abundance and activity of harepopulations will be assessed on a sample of AEfarms and suitable controls during spring/summer2011 and 2012. Juvenile hares will be tagged toenable survival estimates to be calculated and theeffectiveness of the option to be assessed.

Invertebrate speciesSpecific AE scheme options will be sampled insummer 2011 and 2012 to assess the abundanceand diversity of particular invertebrate groupscompared to suitable controls. This will includepitfall trapping of ground-dwelling invertebrateson ungrazed grass margins and surveys of aquaticmacro-invertebrates in waterways associated withenhanced riparian zones. Further research on theutilisation of prescribed AE options by otherinvertebrate groups such as bumblebees andbutterflies may also be undertaken should optionuptake be sufficient.

Plant speciesPlant monitoring will be based on resurveys ofexisting AE scheme habitats from which baselinedata was recorded in 2002/03 (Flexen et al., 2004).This will involve the sampling of permanentquadrats on species-rich grassland, farm woodlandand peatland habitats between May and September2011/12. Changes in species diversity andvegetation composition due to schememanagement over a 9/10 year period will bedetermined. Habitat condition will also be assessedusing specific criteria including vegetationstructure and indicator species.

HedgerowsA survey of hedgerows on a sample of NICMSagreement farms was carried out in summer 2010(Flexen et al., 2011). This included bothestablished hedges and those recently subject torestoration or newly planted. Data were recordedfor a number of attributes including dimensions,integrity and woody species composition. The aimof this baseline survey was to provide informationon the current management and condition ofhedgerows on scheme farms. Condition assessmentof established hedges in terms of structure andconservation value was also been undertaken. Afuture resurvey will assess the effectiveness of

NICMS hedgerow management options anddetermine if the targets for the length and conditionof hedgerows under the scheme have been met.

ConclusionsThis research and monitoring project will assessthe effectiveness of AE schemes in enhancingfarmland biodiversity and delivering on optionspecific targets in Northern Ireland. The resultswill be used to recommend how future AEprogrammes could be refined and improvedthrough modification of existing options or thedevelopment of new options.

Given the considerable public expenditure on agri-environment programmes throughout memberstates, it is important that they are seen to beworking and are attaining their targets. This isespecially the case as increasingly AE schemes arebeing identified as a mechanism for assisting withthe recovery and management of priority speciesand habitats.

Currently, approximately 42% of the agriculturalland area in Northern Ireland is under AE schemeagreement, with a target of 50% by 2013. Thisrepresents a very real opportunity to deliverimportant benefits to farmland biodiversity throughthe implementation of programmes and optionsthat can deliver over a large spatial scale andpotentially at the population level. The currentresearch and monitoring programme is critical tounderstanding how current options are workingand how future AE policy and options can bedefined.

AcknowledgmentsThe authors wish to thank DARD for funding theAE Monitoring Programme. Farmland bird surveysare being carried out in collaboration with theRSPB.

ReferencesDARD (2008). Countryside Management Scheme2007-2013 Information Booklet. DARD.

Flexen, M., Johnston, R.J., Cameron, A. andMcAdam, J.H. (2004). Baseline BiologicalMonitoring of the Countryside ManagementScheme in Northern Ireland 2002-2003. Report toDARD. Queen’s University Belfast.

Flexen, M., O’Mahony, D. and McAdam, J.H.(2011). Monitoring of the Agri-environmentprogramme in Northern Ireland: Hedgerow Survey2010. Report to DARD. AFBI, Belfast.

McEvoy, P.M., Flexen, M. and McAdam, J.H(2006). The Environmentally Sensitive Area (ESA)scheme in Northern Ireland: ten-years of agri-environment monitoring. Biology andEnvironment. Cross J. (ed) Royal Irish Academy,106B: 413-423.

Linking catchment activities with conservationstatus of freshwater pearl mussel populationsas an aid to prioritization of rehabilitationmeasures in Ireland.

E.A. Moorkens1 and S. Downes2

153 Charleville Square, Rathfarnham, Dublin 142 RPS Consulting Engineers, Newtown Centre,Annacotty, County Limerick.Email: emoorkens@eircom.net

IntroductionThe freshwater pearl mussel, Margaritiferamargaritifera, and the Nore pearl mussel,Margaritifera durrovensis are endangeredthroughout their world ranges and are protectedunder Irish and European law. High water qualityand an absence of fine sediment infiltration is ofvital importance in maintaining sustainableMargaritifera populations, as clean, stablesubstrate are required for juvenile mussels tosurvive. Individuals can live to over 100 years ofage, but this longevity can mask long termdeclines, as most populations in the world nowconsist only of adult mussels with no sustainablejuvenile recruitment occurring to replace the agingadults. Juvenile Margaritifera are much moresensitive than adults, as they spend their first fiveyears completely buried in the river bed substrate,where they require high oxygen exchange, acommodity that is lost when river bed substratesbecome clogged through fine sediment infiltrationor eutrophication (Moorkens, 1999).

The Republic of Ireland has carried out wideranging studies which have led to the developmentof 27 sub-basin management plans in order toprotect and restore the Natura 2000 populations ofMargaritifera for which it has responsibility underthe EU Habitats and Species Directive (the 27 draftFreshwater Pearl Mussel Plans can be downloadedfrom WFD Ireland1). The conservation statusvaries greatly among the 27 populations, and theyare approximately equally divided into 5 differentcategories. The status categories were correlatedwith levels of pressures found in the catchmentstudies in order to establish the likely cause andeffects of freshwater pearl mussel decline anddetermine a strategy for measures to lead towardsrecovery.

Materials and MethodsThe 27 populations that belong to the 19 candidateSpecial Areas of Conservation (cSAC) for

1

http://www.wfdireland.ie/docs/5_FreshwaterPearlMusselPlans/Freshwater%20Pearl%20Mussel%20Plans%20March%202010/

Margaritifera cover the species range in theRepublic of Ireland (Figure 1).

Surveys for Margaritifera included a) anassessment of the distribution of mussels withinthe designated river systems, b) an estimate of thenumbers of adult mussels and their densities withineach population, c) an estimate of recruitmentsuccess from size profiles measured from 0.5m x0.5m quadrats at various locations of potentialjuvenile habitat within the population and d) adultmussel counts from permanent repeatable transectsused in the ongoing monitoring of thesepopulations. Methods followed Moorkens (2011).Redox potential differences were measuredbetween the open water and the water within thesubstrate 5cm below the river bed surfaceaccording to the methodology of Geist andAuerswald (2007).

The sub-basin plans utilised GeographicalInformation Systems (GIS) datasets, generatedthrough the various elements and Programmes ofMeasures Studies (PoMs) of the Water FrameworkDirective. Water quality data was obtained fromthe local authorities together with the EPA, andCORINE land cover maps were provided to theRiver Basin District Projects under licenceagreement from the EPA. Livestock Unit Densitymaps were provided by the Department ofAgriculture, Fisheries and Food (DAFF) to theRiver Basin District Projects, to facilitatepreparation of the RBD characterisation reports. Itis based on the CSO data from 2002 and providesthe average LU densities on a DED basis. Whilstthis data set is nine years old, it provides a generalguide to the level of livestock unit density in eachsub basin catchment rather than absolute values ona field by field basis. Forestry coverage detailswere made available through the Western RBD2

Forestry and Acidification PoMs. Finally, locationsand numbers of on-site waste water systems wereassessed for % cover in each of the 27Margaritifera SAC catchments again from theWestern RBD PoMs.

ResultsPopulations were ranked following theMargaritifera surveys into 5 categories withroughly even numbers of populations in eachcategory (Table 1). Catchment characteristics andcatchment land use were analysed by statuscategory. The best remaining populations areassociated with small catchments, with lowintensity land use and with lakes upstream (Table2).

2 http://www.wrbd.ie/index.htm

Figure 1. Map of Ireland showing locations ofMargaritifera cSAC Catchments.

Table 1. Ranking system used to divideMargaritifera SAC Catchments into status groupsfrom 1 (best) to 5 (worst).

Ranking Definition Number ofcatchments

(% ofcatchments)

1 Very large populations of adults(200,000+), all ages of juveniles,some juveniles in more than onearea

5 (18.5%)

2 Large widespread populations ofadults, or smaller numbers in goodbut restricted habitat, somejuveniles in more than one area

5 (18.5%)

3 Large numbers of adults, somedecline from larger numbersevident, few juveniles

6 (22.2%)

4 Small numbers adults fromhistorical evidence (<20,000), veryfew juveniles

5 (18.5%)

5 Very poor population of adults(<10,000), few or no juveniles

6 (22.2%)

The average sizes of the poorest status catchmentsare on average eight times larger than those of thebest status catchments. Low intensity land usecover was estimated from the CORINE database,and in Ireland this includes peat bogs, naturalgrassland, moors and heathland, broad-leavedforests, transitional woodland scrub, inlandmarshes, bare rock and sparsely vegetated areas.

The average percentage of these combinedcategories within a catchment for the highest statuspopulations was 91.85%, compared with only33.85% total for these combined categories inSACs in the worst status category.

Table 2. Status of Margaritifera populationcompared with size of catchment (km²), land use incatchment and presence or absence of a lakeupstream (* = significant at 0.05; ** = significantat 0.01 by status compared with expected mean ofall rivers).

Rank ofpopn.status

No. ofcatchments

Averagesize ofcatch-ment(km²)

Average%extensivelandcover(CORINE)

Number(%) ofcatchmentswith lakesupstream

1 5 72 91.85% 5 (100%)2 5 117 80.85% 5 (100%)3 6 86 77.47% 5 (83%)4 5 158 33.6% 1 (20%)5 6 613 33.85% 1 (17%)Total(mean)

27 (206) (48.31%)(37%)

Sign. ** ** *

The corollary values of 8.15% (for the best statuscatchments) and 66.15% (worst status) were foundfor intensive land use categories, which includedurban land uses, arable and intensive pastures, andconiferous forestry plantations. The top 14 beststatus catchments all have lakes upstream ofMargaritifera habitat.

These results demonstrate that the species isparticularly sensitive to cumulative effects thatlead to stepwise increases in fine sediment and/ornutrients. Further analyses of individual pressuresconfirm that this is so. Table 3 demonstrates that asindividual pressures intensify, pearl mussel statusdrops. The exception is for coniferous forestry,particularly on peat, which is a pressure in the bestpearl mussel catchments as well as the poorest.There is some discrepancy between the CORINEdata and the forestry data, suggesting that theCORINE dataset may be interpreting someconiferous plantations as broad-leaved forestry.

Table 3. Status of Margaritifera populationcompared with numbers of licensed outfalls,numbers of on-site septic systems, % of catchmentwith >1.5 livestock units per hectare size and % ofcatchment (km²) under commercial forestry (* =significant at 0.05; ** = significant at 0.01 bystatus compared with expected mean of all rivers).Rankof pop.status

No. ofon-sitesepticsystemsper km²

Average %of catchmentwith >1.5livestockunits perhectare

% ofcatchment(km²) undercommercialforestry

% ofcatchment(km²) withcommercialforestryplanted onpeat

1 2.01 0 8.7 5.632 3.4 0 10.6 3.713 7 0.04 10.5 4.844 9.5 26.1 18.9 2.575 8.2 50.2 11.8 3.02Total(mean) (6.02) (15.27%) (37%) (3.95)Sign. ** ** *

As the pressures are acting on the river bedsubstrate quality, two important parameters can beused to assess risk from nutrients and from finesediment. These are orthophosphate (measured asunfiltered molybdate reactive phosphorus), whichis normally the limiting factor in filamentous algalgrowth and thus the most important nutrient in theassessment of eutrophication, and redox potentialloss, which is a proxy measure of the loss ofoxygen between the open water and the interstitialhabitat where juvenile mussels live. Table 4 showsthe results of the analysis of these parameters bymussel population status. Orthophosphate levelsrarely exceed the detection limit in the best statusmussel populations, whereas there are detectableand sometimes high levels of phosphorus in thepoorer status rivers. Similarly, loss of redoxpotential gets higher as mussel status declines.

Table 4. Status of Margaritifera populationcompared with mean orthophosphate levels(Source; Local Authorities) for 2005 for riversmeasured (n rivers=18, n samples =361), and lossof redox potential at 5cm (%) (n rivers =16).

Rank ofpopulationstatus

Meanortho-phosphate(mg/l P)

Mean lossof redoxpotential(%)

1 <.005 20.82 <.005 223 0.007 31.84 0.03 31.35 0.03 43

DiscussionAs the main mechanism for managing catchmentprotection in Ireland is the River Basin Districtlevel under the Water Framework Directive, 27sub-basin plans have been prepared using theresources and approach of the larger River BasinCatchment Plans. Using the datasets available, andfrom survey work and walk-over studies, thepressures within each catchment that are negativelyaffecting the pearl mussel populations, or wouldnegatively affect their return to favourableconservation status, can be assessed, and thenmeasures outlined that need to be taken in order toremove these pressures. In order for the measuresto be implemented, they need to be clear in nature,specific in their locations, and the responsibility fortheir implementation identified.

To build on the information distilled during thesub-basin plan project, the specific pressures andtheir extent in catchments, along with the likelymeasures needed to rehabilitate river water andriver bed quality, the estimated timescale ofrehabilitation and the likely timescale of theextinction of the mussel population were allassessed in order to produce a strategy for

prioritisation of effort in Margaritiferaconservation for Ireland (Moorkens, 2010). Thestrategy recognised that there is high cost involvedin the rehabilitation of sustainable Margaritiferahabitat. The required result is that rivers areconsistently at close to reference level, with nosignificant loads of nutrients or fine sedimentsentering the river. For this to be achieved,measures will need to be taken to stop thepressures at source, through de-intensification, oralong the pathway between the source and theriver, for example through drain blockage, fencingand management of buffer zones. Both types ofmeasures are expensive to undertake, andresources to carry them out are limited.

The strategic approach identified eight catchmentsthat would be likely to provide the best return forinvestment in terms of current mussel numbersprotected, and the ability to return populations tosustainable reproduction levels. A further twocatchments were added to cover the Irish range. Ifstrong efforts were made to restore function in thetop eight catchments alone, an estimated 2 millionmussels could be saved from net loss over the nextten years. While the top eight catchments compriseless that 10% in area of the 27 sub-basinspopulations, a total of 92% of the IrishMargaritifera resource (mussel numbers) livethere, so there is immense value to be gained byconcentrating the conservation resource in theseareas. However, only one of these eight catchmentsis currently in favourable conservation status.These catchments are highly sensitive to any formof intensification, so very careful management ofall activities together with restoration measureswill be required in all eight on an ongoing basis.

The future situation may be somewhat worse thanpredicted, as evidenced by the large areas ofconiferous plantation already present in thepopulations that are currently at the highest status,particularly forestry on peat soils. Many of theseareas were planted at a time before water qualityprecautions were considered, and are approachingthe stage where they are ready for their first clearfelling. Unless this can be done in a manner thatcan mitigate negative effects of phosphorus andsediment release, severe damage to thesepopulations will result.

In the case of the conservation of an IUCNendangered species such as the FPM, the value isthe value of its conservation (Unpublished Report,2010). High value conservation areas haveeconomic value that can be directly attributed totheir ecological features, both in terms of theemotional value of knowing that a catchment isoperating in a sustainable manner, and in the wider

ecosystem services that non-intensively managedland and clean rivers provide. Because of the highsensitivity of the freshwater pearl mussel, dilutingconservation effort over small pockets of land inlarge catchments would provide negligibleconservation return. The desired results can onlybe attained through co-operative catchment-wideeffort. Socio-economic models have beendescribed that suit high nature value farming over alandscape scale, but they must benefit both profitand biodiversity (Polasky et al., 2005; Cooke et al.,2009). Polasky (2006) concludes thatconservationists have soft hearts, but that effectiveconservation also requires hard heads. It is difficultfor conservationists to accept that a population of arare species cannot be saved through lack ofresources, but resources are essential to allow areturn to sustainable low intensity land use in thesemost important catchments. Assessing theinternational conservation value of the FPM wouldappear to be an acceptable mechanism forvaluation, as these values are based on the idea ofrarity and the need for protection internationally,rather than just the benefits of the FPM to Ireland.(Unpublished Report, 2010)

Protection for the freshwater pearl mussel has astrong legal basis, and not fulfilling Europeanrequirements under the Habitat’s Directive wouldresult in severe penalties in the form of fines, thusthe cost of not restoring sustainable high qualityriver habitats for Margaritifera is much higherthan ecosystem service loss alone. The field-by-field methodologies of removing risk from juvenilemussels and their habitat has yet to be fullydeveloped, but the analysis of studies on themussel populations themselves and the nature oftheir pressures as described in this paper haveprovided welcome clarity to guide the conservationefforts that need to be taken.

AcknowledgmentsThe Irish sub-basin catchment management planproject was funded by North-South share under theWater Framework Directive. Funding is fromDonegal County Council and the National Parksand Wildlife Service. The project team wasmanaged by RPS Consulting Engineers. Allmembers of the project team are gratefullyacknowledged. Pearl mussel data was from surveywork carried out by Ian Killeen, Eugene Ross andEvelyn Moorkens.

References

Cooke, I.R., Queenborough, S.A., Mattison,E.H.A., Bailey, A.P., Sandars, D.L., Graves, A.R.,Morris, J., Atkinson, P.W., Trawick, P.,Freckleton, R.P., Watkinson, A.R. & Sutherland,

W.J. (2009) Integrating socio-economics andecology: a taxonomy of quantitative methods and areview of their use in agro-ecology. Journal ofApplied Ecology 46, 269–277.

Geist, J. & Auerswald, K. (2007) Physicochemicalstream bed characteristics and recruitment of thefreshwater pearl mussel (Margaritiferamargaritifera). Freshwater Biology 52, 2299–2316.

Moorkens, E.A. (in press) Margaritiferamargaritifera: Stage 1 and Stage 2 surveyguidelines. Irish Wildlife Manuals No. 12, Version2. The National Parks and Wildlife Service,Department of the Environment, Heritage andLocal Government, Dublin, Ireland.

Moorkens, E.A. (2010) Addressing theconservation and rehabilitation of Margaritiferamargaritifera populations in the republic of Irelandwithin the framework of the habitats and speciesdirective. Journal of Conchology 40, 339-350.

Moorkens, E.A. (1999) “ConservationManagement of the Freshwater Pearl MusselMargaritifera margaritifera. Part 1: Biology of thespecies and its present situation in Ireland”. IrishWildlife Manuals No. 8. The National Parks andWildlife Service, Department of the Environment,Heritage and Local Government, Dublin, Ireland.

Unpublished Report, (2010) Assessing benefitvalues of the maintenance and improvement of theFreshwater Pearl Mussel (Margaritifera sp.)populations in Ireland. NS Share project.

Polasky, S. (2006) You can’t always get what youwant: Conservation planning with feedback effectsProceedings of the National Academy of Sciences103, 5245–5246.

Polasky, S., Nelson, E., Lonsdorf, E., Fackler, P. &Starfield, A. (2005) Conserving species in aworking landscape: land use with biological andeconomic objectives. Ecological Applications 15,1387–1401.

The role of habitat creation in the recovery ofthe Irish grey partridge Perdix perdixK. Buckley1, C. O’Gorman2, P. Comerford1, V.

Swan1, C. Flynn1, T. Carnus3, B. Kavanagh4 and

B.J. McMahon3

1NPWS, Dept. of Environment, Heritage and LocalGovernment, Conway Estates, Kildare Town, Co.Kildare, Ireland, 2The British Association forShooting & Conservation BASC, Marford Mill,Rossett, Wrexam UK,3UCD School of Agriculture,Food Science and Veterinary Medicine, UniversityCollege Dublin, Belfield, Dublin 4, Ireland, 4RoyalCollege of Surgeons in Ireland, Medical Universityof Bahrain, Kingdom of BahrainEmail: kieran.buckley@environ.ie

Introduction

The grey partridge (Perdix perdix) is a farmlandbird species which has declined to one remainingnaturally occurring population in the IrishRepublic. It is a red listed species in Ireland (Lynaset al., 2007). The Irish population is now confinedto post industrial peatlands in Boora, Co. Offaly.Management of the species in this landscapecreates suitable habitats for grey partridge,including, nesting, brood-rearing and overwintercover. This habitat creation involves planting of 4meter wide linear strips which included kale,lucerne, triticale mixes, unsprayed cereal,wildflower strips and LINNET plots. Incorporatingthese habitats into modern agriculture is the key tothe future survival of the grey partridge as anaturally occurring species in Ireland. This paperdocuments the role that habitat creation had inincreasing the population from 17 in 1991 to 317in 2010. In addition, we discuss the viability andrequirements for range expansion of the speciespopulation beyond the nucleus in Boora.

Materials and Methods

Since 1991 the last remaining grey partridgepopulation in ROI has been studied on a 150ha siteat Boora, a cutaway peatland. This habitat is notideally suited to grey partridge since this species istypically a lowland farmland bird and classified asa lowland Farmland Indicator species (Gregory etal., 2004).

Conservation measures

The dynamics of post industrial peatland are notideally suited to the management of an iconicfarmland bird like the grey partridge. However,sufficient areas of habitat, including nesting coverin the form of tussocky grass strips, beetle banks,and kale based brood rearing strips wereintroduced in 1999. Habitats to benefit the speciesincluded 100m x 4m wide linear strips of kale,lucernes, triticale mixes, unsprayed cereal

headlands, wildflower strips and LINNET plots.To date a total of 120ha of combined nesting coverand over-wintering cover have been created inBoora, combined with nearly 2km of beetle banks.A consistent predator management programme hasbeen implemented since 1996; predator countscorresponding to total number of predatorseliminated. Population augmentation was carriedout from 2002, but these data are not includedhere.

Data analysis

Autumn population, spring pairs and chick survivalrates (Potts 1986) were analysed using generalisedadditive models that accounted for temporalcorrelations and non-linear effects of annualrainfall. Effects of different habitats on populationvariables were assessed in models that alsoincluded log transformed predator counts. AkaikeInformation Criterion, with a correction factor for afinite samples size (AICc), was used to assess theeffects of management variables on populationvariables following Zuur et al.,(2009).

Results and Discussion

Partridge autumn population and spring pairsincreased over time (Figure 1; P< 0.001). Anincrease in predator counts was significantlyassociated with a decrease in partridge autumncounts (P< 0.01), and chick survival (P<0.001).

Figure 1. Autumn populations of grey partridge

over 20 years. (No data in 2007.)

Cover creation appeared to be the only significanthabitat variable associated with autumn counts.However, overall habitat management may havealso contributed to the increase in partridgepopulation through its effects on chick survival(Table 1).

Table 1. Effects of habitat management variables

on partridge population variables. P values are

given corresponding to AICc comparisons between

models with and without each habitat type.

Beetle

bank

Nesting Cover

Autumn pop. 0.17 0.10 0.05

Spring pairs 0.92 0.69 0.29

Chick survival 0.03 0.10 0.09

Conclusions

Grey partridge population recovery was achievedthrough the multiple influences of predatormanagement and habitat creation in atypicalhabitats for this species. Although multiple factorshelped reverse the decline of grey partridges at thestudy site, appropriate habitat management wasimportant. This indicates that habitat managementfor grey partridges would be suitable in a targetedagri-environment scheme in order to facilitate arange expansion within Irish agriculturalecosystems.It must be stated that single species measures havelimited longevity and success as part of an agri-environment schemes. Therefore, ongoinginvestigations regarding the benefits of specifichabitat measures for partridges on the site in Booraare required to demonstrate, where possible, howother important elements of farmland biodiversitymay also thrive. There is ongoing collection ofdata on granivorous passerines which appear tobenefit for the habitat creation measures for thepartridge. Granivorous passerines have generallydeclined across Europe in recent years (Donald etal., 2006).The implementation of an agri-environmentmeasure that would facilitate the expansion of greypartridges from Boora needs to be highly targetedin terms of region, i.e. in the west Offaly area, andin terms of providing habitat that would benefit thespecies such as linear strips of kale based mixedcrops and beetle banks. The measures implementedcould potentially benefit a wide range of farmlandbiodiversity including granivorous passerines andother Red Listed species of conservation concernsuch as the barn owl (Tyto alba). In addition tothis, a range of invertebrate species couldpotentially benefit from the structure of thehabitats created and it is these populations ofinvertebrates that could provide a vital food sourcefor partridge chicks that would increase survivalrate (Potts, 1986).In conclusion, the results from the Boora projecthave been impressive and have answered thequestion as to whether the Irish population of greypartridges can be saved (Kavanagh, 1998).However, in many respects this is only the end ofthe beginning in terms of saving the species in

Ireland. It would appear that modern agriculturehas driven the species out of its traditionalfarmland habitats in Ireland and for a realconservation success we must endeavour to returnit to its natural habitat. A specific, targeted agri-environmental scheme is the logical approach tofacilitate this decolonisation of Irish farmland.There appears to be sound rationale behind such ameasure as numerous other components ofbiodiversity could also benefit, as outlined above.This approach is in agreement with previousresearch (e.g. Whittingham et al.,2007) whichhighlights the fact that targeted agri-environmentmeasures are more likely to yield beneficial resultsfor biodiversity than those applied uniformly innational schemes.

Acknowledgments

We would like to thank National Parks andWildlife Services of the Dept. of the Environment,Heritage and Local Government. The Irish GreyPartridge Conservation Trust, National Associationof Regional Game Councils & Offaly RegionalGame Council.

References

Donald, P.F., Sanderson, F.J., Burfield, I.J. and vanBommel, F.P.J. (2006) Further evidence ofcontinent-wide impacts of agriculturalintensification on European farmland birds, 1990-2000. Agri. Ecosys. Environ. 116: 189-196.

Gregory, R. D., Noble, D. G. and Custance, J.(2004) The state of play of farmland birds:population trends and conservation status oflowland farmland birds in the United Kingdom.Ibis 146 (Suppl. 2): 1–13.

Kavanagh, B. (1998) Can the Irish Grey Partridgebe saved? A national consideration strategy. GibierFaune Sauvage 15: 533-546.

Lynas. P., Newton S.F. and Robinson, J.A. (2007)The status of birds in Ireland: an analysis ofconservation concern 2008-2013. Irish Birds 8:149-166.

Potts, G.R. (1986) The Partridge: Pesticides,Predation and Conservation. Collins, London.

Whittingham, M. J., Krebs, J. R., Swetnam, R. J.,Vickery, J. A., Wilson, J. D., Freckleton, R. P.,2007. Should conservation strategies considerspatial generality? Farmland birds show regionalnot national patterns of association. Ecol. Lett. 10:25-35.

Zuur A.E., Ieno, E.N., Walker, N.J., Savaliev,A.A., and Smith, G.M. (2009) Mixed effectsmodels and extensions in ecology with R.Springer, New York.

REPS and farmland bird populations: resultsand recommendations from the FarmlandBirds Project

A.S. Copland1 and J. O’Halloran2

1BirdWatch Ireland, Crank House, Banagher, Co.Offaly, 2Biological, Earth and EnvironmentalSciences, Enterprise Centre, Distillery Fields,University College CorkEmail: acopland@birdwatchireland.ie

IntroductionThe joint BirdWatch Ireland - University CollegeCork Farmland Birds Project was established toevaluate the impacts on farmland bird populations(as indicators of biodiversity in the widercountryside) of the Rural Environmental ProtectionScheme (REPS) and offer evidence-basedrecommendations to improve REPS as a tool forthe conservation of birds in farmed landscapes.This study also aimed to develop an ecological(bird/habitat) monitoring methodology for REPS.This paper summarises the results from this project(full details can be found in Copland (2009)).

MethodsBird and habitat data were collected during threesummers (April – June, 2003 – 2005) on 122 farms(61 REPS farms paired with 61 non-REPS farms,pairs were based upon farm location andenterprise) distributed across North-west, Midlandsand South-east, which generally reflect a gradientof farming intensity in Ireland from extensivefarms in the North-west to intensive farms in theSouth-east (see Figure 1). In winter (mid-November – mid-February, 2003/04 and 2004/05),41 farm pairs (a subset of the 61 pairs surveyed inthe summer) were surveyed.

Figure 1. Locations of study sites.

REPS impacts on bird populationsAn assessment of the current impacts of REPS wasundertaken for both breeding and wintering birdspecies. Only the eleven basic measures of REPSwere tested. Due to a lack of baseline biodiversitydata, comparative assessments of breeding andwintering bird populations were made using thebird and habitat data. Further analyses of the datawere undertaken at the farm level to identify keyhabitat predictors of bird occurrence, and thesewere then used to suggest outlines for future agri-environment measures for bird conservation in thewider countryside.

An agri-environment evaluation methodFor the study reported here, key farmland habitattypes are identified (five basic habitat types wereused: improved grassland, other grassland, tillage,built and woody habitat; derived from habitat datacollected during fieldwork). Determination of abiodiversity value for each habitat based on keyindicators (the example used is based on thecollected bird data) that incorporates both theability of the habitat to support individuals as wellas different species was developed, and these wereapplied at the farm scale.

Results and RecommendationsThe three seasons of summer fieldwork surveyed atotal of 3,240.05 hectares across the 122 farms. Onthese, some 31,357 individual birds were recordedfrom 72 different species. In the two winterperiods, a total of 2,178.26 hectares on 82 farmswere surveyed. During winter fieldwork there wasa total of 38,810 individual birds recorded from 66species. The sample size of farms, and number ofbirds included in the analyses reported hererepresent the largest for any study undertaken todate on the impacts of farming activities on birdpopulations in Ireland.

REPS impacts on bird populationsOverall bird species diversity, and individualspecies densities and numbers showed nodifferences between REPS and non-REPS farms.There was one exception of Magpie Pica picaduring winter (which occurred at higher densitieson non-REPS farms). The best predictors of birdoccurrence at the farm level in both summer andwinter were the area of grasslands, field size andamount of fence-lines (typically negativelyassociated with bird numbers), and the amount ofhedgerows with trees, which were positivelyassociated with bird numbers. In summer, theamount of non-cropped (woody, wet and built)habitat was also a good (positive) predictor of birdoccurrences. Farm enterprise type was also asignificant predictor, as was geographical region.

North-west

Midlands

South-east

Increasing both hedgerow densities and the areasof non-cropped habitats would likely benefitoverall bird biodiversity, but measures aiming toachieve this should take cognisance of potentialnegative impacts on certain key species. Also, theregional targeting of agri-environment measures,particularly in Ireland where there are cleargeographical differences in farm habitats andenterprise types, should be considered to maximisethe potential of agri-environment schemes. Moreinformation on the summer data can be found inCopland and O’Halloran (2010b) and for thewinter data in Copland (2009).

An agri-environment evaluation methodA model for simple and rapid biodiversityassessments (SARBAS) was developed (Figure 2).

Figure 2. Method to construct SARBAS model.

The identification of the broad habitat categories isthe only requirement to produce the farm level‘biodiversity score’, and it is intended that thiscould be undertaken by non-biodiversity/taxaspecialists (such as agricultural planners involvedin producing agri-environment plans). The methodoutlined here is designed to be best suited tohorizontal (or broad and shallow) agri-environmentschemes such as REPS. Within such schemes, allagri-environment plans produced could incorporatethis biodiversity score, and all plans should statethat this score should be maintained over the termof the agri-environment agreement. Full details onthis evaluation method can be found in Coplandand O’Halloran, 2010a)

ConclusionsIt is clear from this and other similar studies (e.g.Flynn, 2002), that REPS failed to halt the declinesof bird populations on farmland. Furthermore, thebasic eleven measures within REPS failed to haveany impact on bird populations using farmlandhabitats. It is also clear that an integratedmonitoring and evaluation framework for agri-environment schemes such as REPS is alsorequired. However, REPS and Ireland are not alonein this regard, with many other agri-environment

schemes throughout the EU failing on similarobjectives (Kleijn and Sutherland, 2003).

Had REPS continued, a re-survey of the birdpopulations and habitat composition of the farmsstudied here would have provided very valuabledata on the continuing impacts of the scheme.Nevertheless, the data here indicate a direction forthe operation of successful agri-environment inIreland if broad conservation objectives are to bemet. While such measures would struggle to meetconservation targets for many species listed asBirds of Conservation Concern in Ireland (Lynas etal., 2007), there might be opportunities to addressconservation issues in relation to wider countrysidespecies. However, a properly resourced andintegrated monitoring and evaluation programme isessential to maximise any future impacts.

AcknowledgementsThis work was undertaken as part of the projectMaximising the Biodiversity Impacts of REPS,funded by the Department of Agriculture and Foodthrough the Research Stimulus Fund of theNational Development Plan, The Heritage Counciland the Agriculture Environment StructuresDivision in the Department of Agriculture andFood. Many thanks to all who assisted on theproject: advisory group members, fieldworkers,agricultural planners and the farmers themselvesfor allowing access.

ReferencesCopland, A.S. (2009). Bird Population of LowlandIrish farmland (with special reference to agri-environment measures). Unpublished PhD thesis,University College Cork.Copland, A.S. and O’Halloran, J. (2010a) Simpleand rapid biodiversity assessments (SARBAS): Anevaluation method of Ireland’s agri-EnvironmentScheme. Aspects Appl. Biol. 100: 385-390.Copland, A.S. and O’Halloran, J. (2010b) Agri-environment impacts and opportunities for summerbird communities on lowland Irish farmland.Aspects Appl. Biol. 100: 77-87.Flynn, M. (2002). An Investigation of theRelationship between Avian Biodiversity andHedgerow Management as prescribed under theRural Environment Protection Scheme (REPS).Unpublished PhD thesis, Royal College ofSurgeons in Ireland, Dublin.Kleijn, D. and Sutherland, W.J. (2003) Howeffective are European agri-environment schemesin conserving and promoting biodiversity? J. Appl.Ecol. 40: 947-969.Lynas, P., Newton, S. F., Robinson, J. A. (2007).The status of birds in Ireland: an analysis ofconservation concern 2008-2013. Irish Birds 8:149-167.

Apply these scores at the farm level

Derive a “biodiversity score” for each habitat

Collect biodiversity data in each habitat

Identify key taxa Identify habitats

Posters

Learning to improve agri-environmentalschemes: using experts’ judgements to assessthe environmental effectiveness of the RuralEnvironmental Protection Scheme

D. Bourke1,2,3, I. Kurz1,2, L. Dunne2 and J.A. Finn1

1Environment Research Centre, Teagasc,Johnstown Castle, Wexford, 2Rural EconomyResearch Centre, Teagasc, Athenry, 3Departmentof Botany, Trinity College Dublin.Email: bourkedo@tcd.ie

IntroductionAgri-environment schemes (AES) in the EU offerpayments to farmers in return for undertakingmanagement practices (measures) that maintain,enhance or restore the rural environment. InIreland just over €3.1 billion was spent on REPSbetween 1994 and 2009, with 62,000 farmersparticipating in the scheme when REPS 4 wasclosed in 2009 (DAFF, 2009). These payments arefor the delivery of environmental benefits, but theassessment of the environmental effectiveness(EE) of AES such as REPS has proven difficult.Such assessment is increasingly required to satisfyEU agri-environmental legislation, to demonstratevalue-for-money for taxpayers, and to avoidaccusations of trade distortion. Due to the absenceof quantitative, national-scale, environmentalmonitoring data on the performance of REPS,experts’ judgements were used in this study tolearn about the strengths of the REPS and identifyopportunities for improving the delivery ofenvironmental benefits.

Materials and MethodsClear objectives, verifiable targets, and thecollection of appropriate baseline and monitoringdata are fundamental to assessing how well ascheme is designed and performing. Due to thelack of this information in Ireland to date (Bartoliniet al., 2005), a methodology was devised toestimate the EE of REPS. The methodologycombines; a reduction of complex scheme structureinto assessable elements; experts’ judgement of theperformance of these elements, and; the productionof aggregated judgements on the contribution ofsingle or multiple measures to single or multipleenvironmental objectives.

Table 1 outlines the hierarchical set ofenvironmental objectives derived from evaluationcriteria in the Common Evaluation Questionnaireof the European Commission used in this study(European Commission, 2000). We identified thebasic measures in REPS 2 (RDR 1257/1999) thatcontributed to these objectives, and each measure-objective (M-O) pair was the unit of assessment.

Table 1. Summary of environmental objectives.

Environ.

Category

Objective Description

Soil quality1.1 Reduction of soil erosion

1.2 Prevention & reduction of chemical contamination of soils

Water

quality

2.1 Reduction of agricultural inputs potentially contaminating

water

2.2 Impeding the transport mechanisms (from field surface

or root zone to aquifers) for chemicals (leaching, run-off, erosion)

Species

diversity

4.1 Reduction of agricultural inputs to achieve benefits for

flora & fauna

Habitat

quality

5.1 Conservation of high nature-value habitats on farmed land

5.2 Protection/ enhancement of ecological infrastructure,

including field boundaries/ non-cultivated farmland with habitat

function

5.3 Protection of valuable wetland or aquatic habitats from

leeching, run-off or sediments originating from adjacent farmland

Genetic

diversity

6.1 Conservation of endangered breeds/varieties

Landscape

quality

7.1 Maintain or enhance perceptive/cognitive coherence between

the farmland and the natural/biophysical characteristics of the

zone

7.2 Maintain or enhance perceptive/cognitive differentiation

(homogeneity/diversity) of farmland

Each M-O pair was assessed by the following fivecriteria: strength of cause-and-effect relationship,quality of institutional implementation, farmercompliance, extent of participation, and degree ofspatial targeting (Table 2). Seven Irish agri-environment experts conducted the assessment in aone-day group meeting in which assessment scoreswere explained and justified and experts discusseddifferences among their judgements. Theenvironmental performance of each M-O pair wasestimated using the geometric mean (nth root ofproduct of scores) of the assessment scores fromeach of the criteria, reflecting dependentrelationships among factors in achievingenvironmental effectiveness (Finn et al., 2007,2009).

Table 2: Requirements for highest score (5) to beallocated to the assessment criteria for a M-O pair.Criterion Interpretation of criteria

performanceCause-and-effect If a measure, or group of measures,

would be expected to make a majorcontribution towards achieving theenvironmental objective. Themanagement prescriptions are wholly oralmost wholly appropriate to achieve theagri-environmental objective.

Implementationby institutions

The quality of implementation by theinstitution is high

Implementationby farmers

The measure is implemented wholly oralmost wholly in accordance with themanagement prescriptions

Participation Participation exceeds the level requiredto achieve the expected environmentaleffects of the stated M-O pair.

Targeting Participation in the measure wholly oralmost wholly matches the distributionof the relevant environmental pressure

Results and DiscussionThirty eight M-O pairs were assessed by theexperts. As an overall estimate of environmentaleffectiveness of each M-O pair, the geometricmean of the criteria was calculated and categorised(aggregated) according to individual measures(Table 3) and environmental objectives (Table 4).The average criteria scores for each REPS measureshowed that institutional implementation andcausality showed greatest variation (Table 3). Forexample, M2 and SM3 were given maximumscores for the causality criterion while M3 and M9were assigned low scores. Farmer complianceconsistently received high scores (4.5), and wasjudged by the experts to least affect theenvironmental effectiveness of the scheme. Thetargeting criterion consistently received lowerscores.

Table 3. Average criteria scores for differentmeasures in REPS 2 (maximum score for higheffectiveness = 5). The final column indicates thegeometric mean of the five criteria.

REPSMeasure

Causality

Implementation

Targeting ParticipationGeometric

meanInstit. Farm

M 1 4.7 5.0 4.5 2.3 2.9 3.5

M 2 5.0 3.7 4.7 2.3 2.8 3.4

M 3 1.0 5.0 4.5 2.1 2.8 2.3

M 4 4.0 1.0 4.5 2.3 3.0 2.9

M 5 3.0 5.0 4.5 2.6 3.3 3.1

M 6 0.9 5.0 4.5 2.3 2.7 2.3

M 7 5.0 4.0 4.5 2.8 2.6 3.5

M 8 3.0 5.0 4.5 2.7 2.9 3.1

M 9 1.0 5.0 4.5 2.6 2.0 2.2

M A 4.0 1.8 4.5 2.7 2.9 3.1

SM 3 5.0 5.0 4.5 2.6 3.1 3.7

SM 4 4.4 5.0 4.5 2.6 2.1 3.3

SM 6 5.0 5.0 4.5 2.5 2.2 3.4

The effectiveness of REPS 2 measures in meetingenvironmental objectives is outlined in Table 4.Relatively low scores for participation andtargeting reflect the experts’ belief that the schemecould not achieve its environmental objectiveswithout increased participation by larger, moreintensive farms. The scheme targeted smallerfarms by having a per hectare payment that onlypaid for the first 40 ha. Overall, the basic scheme(Measures 1 to 11) adopts a one-size-fits-allapproach, and incorporates little (if any) regionaldifferentiation. Participation scores for water andsoil quality were slightly lower than those forbiodiversity, probably reflecting the greaterdifficulty in achieving widespread improvementsin water and soil quality. Scores for causality andinstitutional implementation showed much greatervariation, and this reflected specific reasons(deficiencies and good performance) at the level ofindividual M-O pairs (Table 3).

Table 4. Average criteria scores for different EUenvironmental objectives in Ireland’s agri-environment. Data are based on high priority M-Opairs (standard deviation in parentheses).

ConclusionsThe use of experts’ judgements can contribute tolearning how to improve the environmentaleffectiveness of agri-environment schemes. Thisapproach can help identify specific reasons for theineffectiveness of individual measures. It is clearthat verifiable targets, specific to the schemeobjectives, along with relevant monitoring data, arealso required to assess the environmentaleffectiveness of REPS. The experts’ scoreshighlighted many positives in the Irish agri-environment scheme; however, they also identifiedwhere improvements can be targeted to helpincrease environmental delivery.

AcknowledgementsWe thank the experts for their contributions. Thiswork was developed as part of the EU FP6 ITAESproject (SSPE-CT-2003-502070).

ReferencesBartolini, F., Finn, J. A., Kurz, I., Samoggia, A. &Viaggi, D. (2005) Using information from mid-term Evaluations of RDP for the multicriteriaanalysis of agri-environmental schemes. EAAEConference, Copenhagen, August 2005.DAFF. (2009). REPS Fact Sheet 2009. Dept. ofAgriculture, Fisheries, and Food, Dublin.European Commission (2000). Commonevaluation questions with criteria and indicators.Finn, J.A., Kurz, I., and Bourke, D. (2007).Multiple factors control the environmentaleffectiveness of agri-environmental schemes:implications for design and evaluation. Tearmann,6: 45-56.Finn, J.A., Bartolini, F., Bourke, D., Kurz, I., andViaggi, D. (2009). Ex post environmentalevaluation of agri-environmental schemes usingexperts’ judgement and multicriteria analysis. J.Env. Planning and Management, 52: 717-737.

Implementation

Environ. Obj. Causality

Instit. Farm

Targeting Participation n

Soil quality4.0 (2.0) 3.0 (2.3) 4.6 (0.3) 2.4 (0.1) 2.6 (0.2) 4

Water quality3.1 (1.9) 5.0 (0.0) 4.5 (0.0) 2.4 (0.1) 2.6 (0.4) 10

Species diversity2.4 (1.9) 4.2 (1.8) 4.5 (0.0) 2.4 (0.2) 2.6 (0.5) 5

Habitat quality3.3 (2.1) 3.5 (1.9) 4.5 (0.0) 2.5 (0.4) 2.8 (0.4) 14

Genetic diversity5.0 (0.0) 5.0 (0.0) 4.5 (0.0) 2.6 (0.0) 3.1 (0.0) 1

Landscape4.0 (2.0) 4.8 (0.5) 4.5 (0.0) 2.7 (0.1) 2.9 (0.2) 4

Mean3.3 (1.9) 4.1 (1.6) 4.5 (0.1) 2.5 (0.3) 2.7 (0.4) 38

Halting biodiversity loss: the potential of High

Nature Value farming in north-west Ireland

P.G. Boyle1, M. Hayes-Fitzpatrick2, M.Gormally2, J. Moran1

1I.T. Sligo, Ash Lane, Sligo. 2National Universityof Ireland, GalwayEmail: pamgcboyle@gmail.com

IntroductionHigh Nature Value (HNV) farming is lowintensity farming associated with a high diversityof semi-natural habitats and species. Despite thefact that the majority of biodiversity throughoutIreland and the European Union (EU) is formedand managed by agricultural practices, there havebeen very few actions taken to protect thisimportant resource.

A number of projects carried out on an EU scalehave attempted to identify land use and changes inland use (Billeter et al., 2008; Doxa et al., 2010).Most notably, the CORINE project (Bossard et al,2000) has identified an increase of 35% in ‘arableland’ in Ireland between 1990 and 2000; thisincludes land used for grass-based silageproduction. This suggests that potential areas ofHNV farmland continue to be lost as farmingpractices intensify. HNV farming is alsothreatened by land abandonment, as farming nolonger appeals to younger generations or as low-intensity farming becomes uneconomical.

Apart from CORINE, which has recognised scale-related limitations, there has been very little effortdirected towards identification of the extent orquality of HNV in Ireland (Sullivan et al., 2010).This project aims to remedy this by developing amethodology which can be used to identify theextent and quality of HNV in the North West ofIreland, with potential for utilisation throughoutIreland.

The project is composed of two individual studies,the results of which will be amalgamated toprovide a suite of indicators that can be used byagricultural planners, farmers and other interestedparties for the identification and monitoring of theextent and quality of HNV farmland in Ireland.Management recommendations for HNV farmingwill also be produced which may be used tocontribute towards increasing overall biodiversityof farms.

Materials and MethodsThis project will study HNV farming on twolevels; field scale and landscape scale. It is hopedthat by incorporating the results from both strandsthat a standardised methodology for identificationand monitoring of HNV will be produced.

Both study scales are linked strongly by keepingthe farmer and the current farming practices as aconstant consideration throughout. Owners ofstudy farms will be provided with a questionnaireto gather details about land use, fertiliser inputs,livestock units and other factors which mayimpact farmland biodiversity. These details willbe considered on both the field scale andlandscape scale to gain a greater understanding ofwhat management techniques create and maintainHNV farmland in Ireland.

Landscape scaleLandscape scale identification of HNV farmingwill be focused primarily on the use of remotesensing techniques. The project hopes to constructa GIS model which can be used with aerialphotographs, as these are the most accessibleimagery available for agricultural planners, etcwho may use the system in future. This will beground truthed by carrying out plant communityanalysis on selected farms as per Sullivan et al.(2010).

In addition to this, it is envisaged that butterflyand bird data will be collected at a farm level andwill be incorporated into a suite of indicators,which can be utilised for the identification ofHNV farmland.

Field scaleThe focus of the work at field level centres on aninvestigation of the terrestrial invertebrates ofHNV farmland, in landscape ecosystems withdiverse habitats that are typical of the West ofIreland. Diptera and Carabidae will be sampledextensively on selected sites during the samplingseason of 2011. Data will be analysed todetermine how HNV farming practices influenceterrestrial invertebrate and associated plantcommunities. The second year of sampling willinvolve the testing of additional sites usingselected indicator species and the third year of thestudy will see the more broad application of use ofindicators in other regions of the country.

Site SelectionIn order to select a suitable study area forinvestigation in year one, a number of factorswere considered. Previous studies have shown theimportance of soil diversity and land use intensityon the presence/absence of HNV farmland(Sullivan et al., 2010). For this reason, a numberof databases were consulted to select a suitablearea (Table 1).

Table 1. Databases used for site selection.

Source Information obtainedCSO DED, size, location, farmed area statistics, etcGSI GeologyEPA Soils, CORINE land cover and land use changeNPWS Designated AreasOSI Aerial photos, contour maps, Discovery series maps

Progress to dateThe study area for year one has been selectedwithin Co. Mayo (Fig. 1) (Table 2). This area waschosen after consideration of the databasesgathered (Table 1) and is representative of avariety of agricultural systems in North WestIreland. The area incorporates a variety ofecosystem types e.g. upland and coastal areaswhich will provide results which arerepresentative and usable throughout the NorthWest region of Ireland.

Fig. 1. Study area, Co. Mayo (year 1).

Table 2. Agricultural statistics of study area (fromCSO, 2000).Electoral Division Population

(persons)

Total Area

(ha)

Area farmed

(ha)

Total no.

Farms

Croaghmoyle 162 3149 873 31

Burren 292 3717 1105 54

Kilmaclasser 534 2353 1916 72

Derryloughan 579 2846 2022 84

Kilmeena 1445 3613 2644 113

Islandeady 1000 4202 2271 119

Total 4012 19880 10831 473

Future workYear one will focus on gathering baseline data atfield, farm and landscape levels which will beanalysed by the end of the first year. This willallow us to develop and test possible identificationtechniques and models on a wider scale in yeartwo and build on gathered data again in year three.

It is hoped that by the end of this project that thefirst steps towards an easily understood,accessible HNV farmland identification tool orkey will have been developed and tested. This canthen be used by agricultural planners and farmersalike, who will then be able to easily identify andmanage areas of biodiversity importance.

The role of HNV farmland will becomeincreasingly important as the focus of theCommon Agricultural Policy shifts towards‘green’ payments. Ireland has great potential tobenefit from this refocusing. It is thereforeessential that adequate identification andmonitoring methodologies are established by thetime the new policies are implemented tomaximise the gains. This project will buildtowards advancing the status of HNV farmland inIreland, which seems to have generally remainedstatic in recent years. This will contribute toputting Ireland to the fore in terms of sustainableagriculture within Europe.

ReferencesBilleter, R., Liira, J. Bailey D., Bugter, R. et al(2008). Indicators for biodiversity in agriculturallandscapes: a pan-European study. J. Appl. Ecol.45: 141-150

Bossard, M., Feranec, J. and Otahel, J. (2000).CORINE land cover technical guide – Addendum2000. European Environment Agency.

Central Statistics Office. (2000). Census ofAgriculture 2000. http://www.cso.ie/px/

Doxa, A., Bas, Y., Paracchini, M., Pointereau, P.,Terres, J.M. and Jiguet, F. (2010). Low-intensityagriculture increases farmland bird abundances inFrance. J. App. Ecol. 47: 1348-1356

Sullivan, C.A., Sheehy Skeffington, M.,Gormally, M.J. and Finn, J.A. (2010). Theecological status of grasslands on lowlandfarmlands in western Ireland and implications forgrassland classification and nature valueassessment. Biol. Cons. 143: 1529-1539

Upland commonages in Connemara – canbiodiversity be used to inform sustainable agri-environmental policy?

B. Canning1, A. Bell1,2, B. Mongan2, M.Gormally1, M. Sheehy Skeffington2 and T. vanRensburg3

1Applied Ecology Unit, NUI Galway, 2Botany &Plant Science, NUI Galway, 3Department ofEconomics, NUI GalwayEmail: brendancanning1@hotmail.com

IntroductionWhile the Connemara landscape has been heavilyinfluenced by agriculture for some time, the effectsof modern farming practices in recent decadeshave been much more rapid and intense. TheEuropean Less Favoured Areas rural developmentpolicy introduced headage payments in the early‘90s which contributed to overgrazing of uplandcommonages. To address environmental problems,several policy initiatives were introduced. Theseincluded the EU policy of decoupling, the IrishRural Environmental Protection Scheme which isan Irish interpretation of EU policy and theCommonage Framework Plans which are anational policy. The effects of these measures onbiodiversity in the uplands have not been studied inany great detail. This research investigates thebiodiversity value of upland commonages inConnemara, with particular reference to vegetationand ground beetle (Carabidae) communities asindicators of habitat condition. This study, whichforms an intrinsic part of an extensive socio-economic investigation of Connemara, willcontribute to the development of agri-environmental policy in supporting the provisionof agrobiodiversity on farms in the uplands.

Materials and MethodsHabitat maps of 16 upland commonages inConnemara were produced using Fossitt (2000). Ahabitat condition score (good, moderate and poorcondition) was also given to each distinct peatlandhabitat, based on guidelines produced by theNPWS - then called Duchás (Anon 1999). In 2010,thirteen wet heath habitat sites of different habitatcondition were selected from six of thesecommonages. Sampling was done in wet heath asit was one of the dominant habitat types in theregion and it showed the greatest variation incondition of all habitats. Data were collected from6 poor, 4 moderate and 3 good condition sites.Eight 1m2 vegetation relevés were recorded at eachof these sites giving a total sample of 104 relevés.Each plant species was recorded and its abundancemeasured as percentage cover. As well ascollecting data on ground flora, physicalenvironmental data such as vegetation height and

percentage cover of bare ground, slope and aspectwere also recorded. In addition, 3 pitfall traps wereset to capture ground beetles at each site from Mayuntil October 2010, giving a total of 39 traps.These were arranged in a triangle 1.5m apart andcollected fortnightly, giving a total of 390 samples.

Results and DiscussionNMS (Non-metric Multidimensional Scaling) is anon-parametric ordination which does not assumelinear relationships between data (Mc Cune andGrace 2005). The angles and length of theradiating lines indicate the direction and strengthof relationships of the variables with the ordinationscores. The sites in poor condition lie at the topand right-hand side of the ordination (Fig. 1), thosein good to the bottom of axis 3 and those inmoderate condition sites in roughly an intermediateposition. Using the Sørensen index as a distancemeasure, axis 3 represents 52% of the variation inthe matrix while axis 1 and axis 2 (axis 2 notshown) represent 23% and 11% respectively.Percentage bare ground is positively correlatedwith Axis 1 (r = 0.750) whereas percentage shrubcover is negatively correlated with this axis (r = -0.760). Average vegetation height is negativelycorrelated with axis 3 (r = -0.496).

%Bare Ground%Shrub

AvgVGHT

Axis 1

Axi

s3

Site12345678910111213

Fig. 1. NMS ordination showing good (green),moderate (blue) and poor (red) condition of siteswith numbers representing each of the 13 sites.

Ground beetle assemblages also differ according tohabitat condition (Fig. 2). Sites in poor conditionplotted to the left (along with % bare ground),while the sites in good condition plotted to theright and bottom along with average vegetationheight, % forbs and %bryophyte cover. Axis 2represents 33% of the variation in the matrix whileaxis 3 and axis 1 (axis 1 not shown) represent 25%and 22% respectively. Percentage bare ground isnegatively correlated with axis 2 (r = -0.657).Percentage forb (r = 0.569) and bryophyte cover (r= 0.499) were positively correlated with axis 2.

% Bare Ground

% Forbs

% Bryophyte

AvgVGHT

Axis 2

Axi

s3

Site12345678910111213

Fig. 2. NMS ordination of Carabidae showinggood (green), moderate (blue) and poor (red)condition sites with numbers representing each ofthe 13 sites.

Average vegetation height is negatively correlatedwith axis 3 (r = -0.610).There were very significantdifferences between % bare ground, % shrub coverand average vegetation height across all three siteconditions (Table 1). This means that they may beuseful as indicators of site condition. Percentageforb and % bryophyte cover showed significantdifferences between good and poor, and moderateand poor condition sites, but no significantdifference between good and moderate conditionsites.

Table 1. Comparison of sites in different conditionstatus: good and poor (G-P), good and moderate(G-M) and moderate and poor condition (M-P).Significant differences (Mann-Whitney test) arehighlighted in bold.

Carabus clatratus is a species of ground beetle indecline in western Europe as its preferred habitats(natural bogs, swamps and mires) are increasinglydisappearing due to drainage(http://www.habitas.org.uk/groundbeetles/).Ireland is one of the last strongholds for thisspecies and as such, it is a species of conservationconcern. Greater numbers of C. clatratus werefound on poor condition sites (Fig. 3). Thissuggests that agri-environmental policy solelydesigned to improve the condition of uplandcommonages may have a negative impact on thisimportant species.

PMG

18

16

14

12

10

8

6

4

2

0

Site condition

Nu

mb

er

of

ind

ivid

ua

ls

Carabus clatratus

Fig. 3. Boxplot of C. clatratus abundance in sitesin different condition status (G= good, M =medium, P = poor). The median value is shown asa thick black line while the box covers theinterquartile range.

ConclusionsThe plant and ground beetle communities areinfluenced by differences in the condition ofupland commonages. The vegetation and groundbeetle community data show a clear response tohabitat condition, but the ground beetle communitydata responded to different environmentalvariables. Hence, it seems that looking atvegetation alone in assessing habitat condition isinsufficient.There is evidence of degradation throughout muchof the upland commonages of Connemara whichmust be addressed through agri-environmentalpolicy. However, it would seem that even the mostdegraded of sites may have some ecological valuefor invertebrates, which is an importantconsideration to have when designing agri-environmental policies.

AcknowledgmentsThe authors would like to gratefully acknowledgethe financial support from the Department ofAgriculture, Fisheries and Food under the NationalDevelopment Plan 2006 Research Stimulus Fund.They would also like to thank the farmers ofConnemara for their kind co-operation andCaitriona Maher for help with the data analysis.

ReferencesFossitt, J. (2000) A Guide to the Habitats inIreland. The Heritage Council, Kilkenny, Ireland.Anon 1999. A Manual for the Production ofGrazing Impact Assessments in Upland andPeatland Habitats. Dúchas – The Heritage Serviceand the Department of Agriculture and Food,Dublin, Ireland.McCune, B., Grace, J.B. (2002) Analysis ofecological communities. MjM Sofware Design,Oregon.http://www.habitas.org.uk/groundbeetles/

%BareGround %Forbs %Shrub

%Bryophytes

AvgVeghgt

G-P <0.001 <0.001 <0.001 <0.001 <0.001

G-M <0.001 0.479 0.003 0.134 0.005

M-P <0.001 <0.001 0.001 <0.001 0.003

Probability of pollinator occurrence increaseswith duration of AES participation and floraldiversity.

T. Carnus1, A. Anderson1, J. Breen2, B.J.McMahon1, V. Santorum2, H. Sheridan1 and G.Purvis1

1School of Agriculture, Food Science andVeterinary Medicine, University College Dublin,Belfield, Dublin 4, 2Department of Life Sciences,University of LimerickEmail: tim.carnus@gmail.com

IntroductionAgricultural intensification and specialisation hasbeen implicated in the loss of habitat and floralresources, resulting in the considerable decline inhymenopteran pollinators over the last sixty years(Goulson et al., 2008). In view of the principal rolethat these organisms play in providing pollinationservices, this trend is alarming. In response toagricultural impacts on biodiversity, the RuralEnvironment Protection Scheme (REPS) has beenimplemented in Ireland in various forms since1994. One particular aspect of this Agri-Environment Scheme (AES) focuses onsafeguarding grassland margins from the effects ofpasture management and nutrient enrichment. Theprescribed exclusion of an area within 1.5m of themargin from grazing cattle, slurry application andsynthetic fertiliser application is expected toimprove floral diversity. However, the positiverepercussions of these measures on field marginflora is likely to take time because soil to plantprocesses that regulate response to nutrient loadinggenerally operate at medium time scales (Musterset al., 2009). The resulting increase in floraldiversity may consequently improve both thehabitat and the food resource necessary tomaintain/improve arthropod biodiversity andecosystem services, and specifically, pollinatorcommunities (Isaacs et al., 2009). Here we presentdata from 119 farms surveyed for pollinators usingpan traps along field margins also surveyed forplant.

Materials and MethodsPastoral farms from thirty distinct 16 km2 squaresof identical soil type were surveyed in 2007 and2008 across three regions (Sligo-Leitrim, Offaly,Cork). Four farms were sampled from each square.They were managed for either dairy (36), beef(32), suckler (44) or mixed (7) production.

Sampling protocolThree pan traps per farm were placed 1m aboveground along a field margin and collected after 48hours during July–August 2007 (60 sites) and 2008(59 sites). All collected pollinators were sorted and

identified to species level. Plant communities weresurveyed in four quadrats each in the margin,within 1m of margin and 20m in field, using BraunBlanquet cover system. The species richness offlowering plants was used as a measure of potentialfloral resources.

Data analysisWe used GLMMs to assess the relationshipbetween pollinator presence/absence (ie binaryresponse variable assessing for the presence of anypollinator), abundance and richness, and REPSparticipation, farm management, and floral plantrichness. Random effects for square were specifiedto account for potential correlation between farmsin the same 16km2 square. We also ran the analysisexcluding squares that did not contain any REPSparticipating farms (4) or only REPS participatingfarms (5). Results did not qualitatively differ, sowe present data collected from all farms.

Table 1. Likelihood ratio test p-values. The effectsof region (Cork, Offaly-Laois, Sligo Leitrim),length of time individual farms have been a REPSparticipant (Time in REPS, in years) and plantspecies richness (Plant richness) on pollinatorpresence-absence (Pre/abs), pollinator richness(Richness) and pollinator abundance (Abundance)were assessed by likelihood ratio test.Response Region

(2df)

TimeinREPS

(1df)

Floralplantrichness

(1df)

Pre/abs <0.01 0.03 0.04Richness <0.01 0.54 0.01Abundance <0.01 0.54 0.09

Results and DiscussionA total of 367 pollinating Hymenoptera (Aculeata)in 42 species were collected. Although norelationship was found between pollinators andfarming system or whether a farm was in the REPSor not (p > 0.1), the probability of pollinatoroccurrence was positively correlated with theduration of REPS participation and richness andabundance with floral resources (Table 1 andFigure 1).

ConclusionsIt may take considerable time for passive AES tohave the desired positive effects on specificelements of biodiversity. Although non-targetorganisms can benefit from general non-targetedmeasures such as those implemented in the REPS,this may be due more to chance than design. Theimprovement of biodiversity status in field marginsmay be greatly enhanced through more activeapproaches. Indeed, pollinators would be morelikely to respond positively to targeted measures

that specifically provide the required habitat andfood resource, as has been found in arablelandscapes (Pywell et al., 2006). These effects arenot only relevant to the conservation of pollinatorsbut also to the wider range of arthropod-mediatedecosystem services, including pest control andfood resources for higher trophic levels. Byproviding pollen and nectar resources, suchtargeted measures improve the chances ofmaintaining sustainable populations of theorganisms that provide these services. Thelivelihood of rural communities is increasinglydependent on production-decoupled subsidy.Delivery of the expected benefits from AEStherefore depends on their design.

Fig. 1. Probability of bee presence in relation tolength of time in REPS. Average probability acrossfarms and 95% confidence intervals are given.

Although the analysis of pollinator populationswas limited by sampling methodology, we givequalitative indication of pollinator distributions onIrish pastoral farms and implicate floral resourcesas the link between farm management andpollinator diversity.

AcknowledgmentsThe authors thank DAFF for the provision offunding under the Research Stimulus Fund 2006.

ReferencesGoulson, D., Lye, G.C., and Darvill, B. (2008)Decline and conservation of bumble bees. AnnualReview of Entomology, 53: 191-208.

Isaacs, R. Tuell, J., Fiedler, A., Gardiner, M., andLandis. D. (2009) Maximizing arthropod-mediatedecosystem services in agricultural landscapes: therole of native plants. Frontiers in Ecology and theEnvironment, 7: 196–203.

Musters, C.J., van Alebeek, F., Geers, R.H.,Korevaar, H., Visser, A. and de Snoo, G.R. (2009)Development of biodiversity in field marginsrecently taken out of production and adjacent ditch

banks in arable areas. Agriculture, Ecosystems &Environment 129: 131-139.

Pywell, R.F., Warman, E.A., Hulmes, L., Hulmes,S., Nuttall, P., Sparks, T.H., Critchley, C.N.R. andSherwood, A. (2006) Effectiveness of new agri-environment schemes in providing foragingresources for bumblebees in intensively farmedlandscapes. Biol. Cons. 129: 192-206.

Positive effects of sown evenness on swardbiomass across three years, nitrogenapplication and cutting intensity.

T. Carnus1, 2, J.A., Finn2, A. Cuddihy2, L. Kirwan3

and J. Connolly1

1Environmental and Ecological Modelling Group,School of Mathematical Science, UniversityCollege Dublin, Belfield, Dublin 42Teagasc, Johnstown CastleEmail: tim.carnus@gmail.com

IntroductionEnvironmental, economic and legal constraints areputting increasing pressure on agricultural systemsto reduce inputs while maintaining or increasingproductivity. This study is concerned with theeffects of diversity of two grass and two cloverspecies on biomass production, in mixed grasslandsystems, under two different levels of nitrogen (N)addition and two levels of cutting height over threeyears.

Materials and MethodsExperimental designA split-plot field experiment was established in2006 at Teagasc Johnstown Castle research centre.The main plot treatment involved varying plantdiversity and the split-plot treatment was a two-way factorial of low and high levels of N addition(~50 kg ha−1 y−1 and ~200 kg ha−1 y−1) and cuttingintensity (~2 cm and ~7 cm). Seed biomassproportions of Lolium perenne, Phleum pratense,Trifolium pratense and Trifolium repens weresystematically manipulated to vary evenness (E, ameasure of relative abundance distribution)according to a simplex design (Kirwan et al.,2007). The design consisted of four monocultures(E = 0), six two-species mixtures (E = 0.67) and 18four-species mixtures dominated in turn by eachspecies (88:4:4:4, E = 0.29 and 70:10:10:10, E =0.64), by pairs of species (40:40:10:10, E = 0.88)and equally represented at the centroid(25:25:25:25, E = 1). The design was repeated attwo levels of overall initial abundance giving atotal of 56 main plots and 224 split-plots. Plotswere harvested three times in 2007, twice in 2008and three times in 2009.

Data analysisTotal yield and unsown species were analysed as afunction of species identity, evenness andtreatment factors (Kirwan et al., 2009) using linearmixed models to account for the split-plot randomeffects and repeated measures over three years.

Results and DiscussionOver three years, both evenness and N additionpositively affected total yield (Fig. 1, Table 1),while cutting decreased total yield.

Table 1. Predicted yields (t/ha) (s.e.) ofmonocultures, and effects of evenness, N andcutting (cut) over 3 years. Figures in bold aresignificant at p < 0.01.

Model terms 2007t/ha

2008t/ha

2009t/ha

Monocultures:L. perenne 7.73 9.02 9.32P. pratense 8.67 11.7 10.5T. pratense 9.68 8.6 9.55T. repens 7.83 8.46 11.1

s.e 0.68 0.67 0.7Treatments

Sowing density -0.34 (0.17) -0.06 (0.17) -0.23 (0.18)Cutting -1.11 (0.47) -0.39 (0.46) -1.48 (0.51)Grass x N 2.47 (0.53) 1.63 (0.53) 2.46 (0.53)Legume x N 0.73 (0.53) 0.33 (0.53) 1.94 (0.58)

Mixture effects:Evenness 2.49 (0.67) 1.72 (0.67) 0.81 (0.70)Evenness x cut -0.47 (0.68) 0.64 (0.67) -0.4 (0.74)Evenness x N -0.88 (0.68) -0.47 (0.67) 0.38 (0.74)

Evenness x N x cut 0.66 (0.96) 0.25 (0.95) 0.01 (1.05)

In the first two years, yield of dry matter waslinearly related to evenness at both N levels andcutting (no evenness by treatment interactions),and was maximum at the centroid (overyielding of1.67 t/ha−1 averaged over 3 years, and Table 1).Over the three years, centroid mixtures produced23% and 14% more biomass than the averagemonocultures, when fertilised with 50 or 200 kgha−1 y−1 of N, respectively.

Fig. 1. Total yield as function of evenness acrossthree years, two cutting heights and two levels ofN. Regression lines indicate overyielding, whilethe intercept indicates average monoculture yields.Dashed regression lines and crosses for the higherlevel of nitrogen; solid lines and circles for lowerlevel.

Due to overyielding, mixture yields at low levelsof N were comparable to the yields of the best-performing monoculture at high levels of N (L.perenne in 2007, P. pratense in 2008 and T. repensin 2009). Overyielding at both high and low Nsuggests that the diversity effect was not solely dueto symbiotic N-fixation.

Fig. 2. Unsown species (weed) biomass as functionof evenness. Regression line shows the quadraticunderyielding in mixtures compared to averagemonocultures (intercept of regression line). Datawere log transformed in analysis.

Biomass of unsown species (weeds) wassignificantly lower in mixtures than monocultures(Table 2 and Fig. 2). Although averaged over threeyears there were no significant differences inunsown species biomass between monocultures ofdifferent species (Table 2), year by year significantdifferences were evident. P. pratense had the mostresistance to unsown species incidence over allthree years, while both clover monocultures hadhigh incidence of unsown species, particularly in2009 where T. repens monocultures werecomposed in some cases of up to 99% unsownspecies.

Table 2. Model fitting for total biomass andunsown species biomass averaged over three years.Likelihood ratio tests were used to assess thesignificance of all effects. Significant effects (p <0.05) are in bold.

Effect tested df Totalbiomass

Unsown sp. biomass

χ2df P(>

χ2df )

χ2df P(>

χ2df )

Sowing density 1 1.41 0.23 0.02 0.88Monocultures 3 8.41 0.04 5.42 0.14Evenness 1 9.95 <0.01 11.1 <0.01N 1 58.5 <0.01 0.23 0.63N x mono. 3 7.55 0.06 0.86 0.83N x evenness 1 0.15 0.7 0.98 0.32Cutting 1 88.6 <0.01 1.8 0.18Cut x mono. 3 1.33 0.72 0.42 0.94Cut x evenness 1 0.05 0.83 0.77 0.38

ConclusionsFour-species mixtures consistently produced morebiomass than the average of the constituent speciesmonocultures. The extent of overyielding declinedin the final year, most likely due to severedeterioration of sown proportions and the veryhigh incidence of unsown species in monocultures,particularly in the clover monocultures (highestyielding monoculture in 2009 was that of T.repens: this was in fact 99% unsown species).The addition of nitrogen had a positive effect ongrass yields (and clover in final year – againprobably on unsown species as opposed to clover)and cutting had an overall negative effect onbiomass. Thus, overall, evenness effects wereunaffected (no interactions between either of thetreatment levels and evenness, which suggests thatthe positive effects of mixtures were relativelyrobust.

AcknowledgmentsT. Carnus was supported by the Teagasc WalshFellowship Scheme.

ReferencesKirwan, L., Connolly, J., Finn, J.A., Brophy, C.,Lüscher, A., Nyfeler, D. & Sebastià, M.T. (2009)Diversity-interaction modelling: estimatingcontributions of species identities and interactionsto ecosystem function. Ecology 90: 2032-2038.

Kirwan et al. (2007) Evenness drives consistentdiversity effects in an intensive grassland systemacross 28 European sites. Journal of Ecology95:530-539.

Conserving Ireland’s farmland birds:performance and prospects of agri-environment schemes

A.S. Copland1, O. Crowe2, J. O’Halloran3 andA.W. Lauder2

1BirdWatch Ireland, Crank House, Banagher, Co.Offaly, 2BirdWatch Ireland, Unit 20 Block D,Bullford Business Campus, Kilcoole, Co. Wicklow,3School of Biological, Earth and EnvironmentalSciences, Enterprise Centre, Distillery Fields,University College CorkEmail: acopland@birdwatchireland.ie

IntroductionUnder the EU Rural Development Regulation(European Union, 2005), Member States arerequired to operate agri-environmental schemes(AES). The stated aim of AES is the “protectionand improvement of the environment”, of whichbiodiversity is identified as a key issue. The RuralEnvironment Protection Scheme (REPS) wasIreland’s AES, operated between 1995 and 2009.The Agri-Environment Options Scheme (AEOS)was launched in 2010 to replace REPS. This papersummarises the current status of farmland birdpopulations and the impacts of AES on these inIreland, using data from recent studies. It exploresthe potential of AEOS in meeting biodiversityobjectives. Bird populations are used as indicatorsof biodiversity since they satisfy many of thecriteria of effective indicators, and have beenwidely used within agricultural ecosystems(Gregory et al. 2005; Purvis et al., 2005).

Farmland Birds in Ireland - Current StatusAnalysis of data from the Countryside Bird Survey(CBS) between 1998 and 2008 (Crowe et al.2010) indicated that agricultural intensificationwas most likely responsible for continued declinesin some farmland bird species. Of three speciesthat showed significant declines over theassessment period, the declines of two (KestrelFalco tinnunculus and Skylark Alauda arvensis)were directly attributed to changes in agriculturalpractices. The analysis also suggested that landabandonment, with consequent decline in farmingactivity, will impact on grassland specialists,including several of conservation concern.

An assessment of the conservation status of allbirds in Ireland in 2007 (Lynas et al., 2007)highlighted a 90% decline in YellowhammerEmberiza citrinella over the last two decades andstated that this species represented the fortunes ofmany other lowland farmland birds that havedeclined in Ireland and across Europe. Theassessment recommended that targeted habitatmanagement for these species, for example

through agri-environment schemes, should givethem the best chance of recovering. Of the 19species listed on the Red List of the Birds ofConservation Concern in Ireland (BoCCI) due tobreeding population concerns, nine (Grey PartridgePerdix perdix, Quail Coturnix coturnix, CorncrakeCrex crex, Lapwing Vanellus vanellus, CurlewNumenius arquata, Redshank Tringa totanus, BarnOwl Tyto alba, Twite Carduelis flavirostris andYellowhammer) are dependent upon farmlandhabitats at some point during the course of theyear. In addition to this, several other bird speciesthat are dependent upon agricultural habitatsappear on the BoCCI Amber List. The only twospecies (Corncrake and Curlew) on the IUCN RedList (species of global conservation concern)breeding in Ireland are both species associatedwith lowland farmland (IUCN, 2010). Also, themost recent regular breeding species to becomeextinct in Ireland, Corn Bunting Emberizacalandra, was a specialist lowland farmland bird,with breeding last recorded in 1992 (Taylor andO’Halloran, 2002).

Impacts of REPS on farmland bird populationsSeveral studies (e.g. Flynn, 2002; Copland, 2009)concluded that REPS had little or no demonstrableeffect on bird populations. Although there is a lackof baseline data against which to compare theimpact of REPS, comparative studies of bothbreeding and wintering bird populations havefailed to show any significant differences betweenbird population on farms undertaking REPSmeasures and those not participating in the scheme.Given the nature of REPS, with its ‘broad andshallow’ approach, the lack of any impact onbiodiversity is to be expected (Vickery et al.,2004). Other benefits, such as improvements towater quality or landscape, are not considered here.

Species-specific AES in IrelandIn addition to REPS, two agri-environmentschemes have operated in Ireland over a similartime-frame. These are the Corncrake GrantScheme (CGS) in core Corncrake areas since theearly 1990s (Donaghy, 2007), and Grey Partridgeconservation at Boora, Co. Offaly since 1991(Copland and Buckley, 2010). The high level ofinput to both projects by specialist staff hasresulted in a stabilisation of Corncrake populationsin Ireland (against a background of huge declinesfrom a national population of thousands as recentlyas the 1970s (Donaghy, 2007)), and an increase inthe Grey Partridge population at the onlyremaining core site (see Figure 1) (see alsocontribution by Buckley et al.). Additionally, bothprojects have delivered benefits for non-targetspecies. Together, these two schemes demonstrate

that an agri-environment approach can work toconserve farmland birds in Ireland.

0

100

200

300

400

500

600

700

800

900

1000

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

Year

Nu

mb

er

of

bir

ds

Grey Partridge

Corncrake

Fig. 1. Number of singing male Corncrakes andautumn Grey Partridge population estimates inIreland, 1991-2010.

Potential impacts of AEOS on biodiversityAs with the two projects above, it has been shownthat AES that set objectives that are focused ondelivering for target species can deliverconservation benefits (Aebischer et al., 2000).Although at present the objectives within AEOSare likely too broad to deliver demonstrableconservation benefits, the overall architecture ofthe scheme (setting an objective for each applicantwho then selects a range of options to meet thatobjective) provides AEOS with the potential toaddress the decline of farmland bird populations inIreland.

In addition to expanding the existing Corncrakeand Grey Partridge programmes, other potentialoptions within AEOS could target breeding waderspecies on lowland wet grasslands habitats acrossIreland, Twite breeding and/or wintering in thenorth-west of Ireland or Chough Pyrrhocoraxpyrrhocorax on coastal grasslands. Non-avianspecies that might benefit from such targetedmeasures could include Natterjack Toad Bufocalamita (building upon the existing NPWSscheme in Co. Kerry) or Great Yellow BumblebeeBombus distinguendus.

If AEOS is to fully realise this potential, morefocused objectives are required. These need toallow species targeting, both geographically withinIreland and for specific habitat types. The currentBird Atlas (2007-2011) will soon be completed,and an update on the distribution of all bird speciesin Ireland will be available. Concurrently, work isongoing to map the distribution of many non-aviantaxa that are likely to be of conservation concern.An opportunity now exists for Irish AES to deliver

on reversing declines in biodiversity in the widerIrish countryside as well as the populations ofmany species of conservation concern, such asfarmland birds.

ReferencesAebischer et al. (2000) From science to recovery:four case studies of how research has beentranslated into conservation action in the UK. In:Aebischer et al. (eds.) Ecology and Conservationof Lowland Farmland Birds. BOU, Tring,Hertfordshire.

Copland, A.S. (2009) Bird Population of LowlandIrish farmland (with special reference to agri-environment measures). Unpublished PhD thesis,University College Cork.

Copland, A.S. and Buckley, K. (2010) Back fromthe brink: Grey Partridge rescued from oblivion.Wings 59: 17-19.

Crowe et al. (2010) Population trends ofwidespread breeding birds in the Republic ofIreland 1998-2008. Bird Study 57: 267-280.

Donaghy, A. (2007) Management of habitats onthe Shannon Callows with special reference totheir suitability for corncrake Crex crex.Unpublished PhD thesis, University College Cork.

European Union (2005) Council Regulation (EC)No. 1698/2005 on support for rural developmentby the European Agricultural Fund (EAFRD).Official Journal of the European Union I 277/1.

Flynn, M. (2002) An Investigation of theRelationship between Avian Biodiversity andHedgerow Management as prescribed under theRural Environment Protection Scheme (REPS).Unpublished PhD thesis, Royal College ofSurgeons in Ireland, Dublin.

Gregory et al. (2005) Developing indicators forEuropean birds. Philos. T. R. Soc. B 360: 269-288.

IUCN. (2010). IUCN Red List of ThreatenedSpecies (http://www.iucnredlist.org).

Lynas, P., Newton, S.F., Robinson, J.A. (2007).The status of birds in Ireland: an analysis ofconservation concern 2008-2013. Irish Birds 8:149-167.

Purvis et al. (2005). The significance ofbiodiversity in agriculture: relevance, aims andprogress of the Ag-Biota Project. Tearmann 4: 29-50.

Taylor, A.J. and O’Halloran, J. (2002). The declineof the Corn Bunting, Milaria calandra, in theRepublic of Ireland. Biol. Environ. 102B: 165-175

Vickery et al. (2004) The role of agri-environmentschemes and farm management practices inreversing the decline of farmland birds in England.Biol. Conserv. 119: 19-39.

The food versus fuel debate – what effect willreplacing traditional crops with Miscanthus xgiganteus have on farmland biodiversity?

M. Emmerson1, D. Bourke2, J. Dauber3, E.O’Rourke4, D. Stanley2, R. Thompson4, P.Whelan4, and J. Stout2.1School of Biological Sciences, Queens UniversityBelfast, 2Department of Botany, School of NaturalSciences, Trinity College Dublin 3Institute ofBiodiversity, Johann Heinrich von Thünen-Institute (vTI), Federal Research Institute forRural Areas, Forestry and Fisheries, Germany.4School of Biology, Earth and EnvironmentalSciences, Environmental Research Institute,University College Cork.Email: bourkedo@tcd.ie

IntroductionThe adoption of the Kyoto Protocol of the UNFramework Convention on Climate Change hasstimulated the search for methods to reduce netCO2 emissions to the atmosphere. The urgency formitigation actions in response to this hasstimulated policy makers to encourage the rapidexpansion of the bioenergy sector, resulting inmajor land-use changes over short timescales.Despite the potential impacts on biodiversity andthe environment, scientific concerns about large-scale bioenergy production have only recentlybeen given adequate attention (Dauber et al.,2010). The SIMBIOSYS (www.simbiosys.ie)project was set up to investigate the impacts of arange of sectors on biodiversity and ecosystemservices, with part of the project’s focus on thosemeasures that may help mitigate the effects ofclimate change. In this paper we therefore aim toassess the impact of growing Miscanthus xgiganteus on former grassland and tillage crops onplant, pollinator and carabid beetle diversity andabundance and the composition of theircommunities.

Materials and MethodsFifty sites were selected across the south east ofIreland, consisting of 10 replicates of 5 treatments(crop types): Miscanthus planted on former tillage(MT), Miscanthus planted on former grassland(MG), Oilseed rape (OS), Tillage control (winterwheat) (CT) and Grassland control (CG) (Figure1). Plants, pollinators (syrphids, bumblebees,solitary bees) and carabid beetles were surveyedat the margin, edge and centre of each site on twooccasions during the summer of 2009. Plants weresurveyed with 1m x 1m quadrats using thepercentage cover abundance scale. Pollinatordiversity and abundance was measured using blue,white and yellow coloured UV pan traps. Carabid

beetle diversity and abundance was measuredusing pitfall traps. Generalised linear models(with either a poisson error distribution or log10transformation where data were not normal) wereused to examine the effect of crop type on speciesrichness and abundance. A generalised leastsquare model was used where variance was nothomogenous. The effect of crop type oncommunity composition was assessed using aPERMANOVA analysis using a Bray CurtisSimilarity matrix. A posteriori pairwise tests wereperformed between crop types. All statistics wereperformed using R (R Development Core Team,2008) and PRIMER v6 (Clarke & Gorley, 2006).

Figure 1. Location of the 50 sites across the southeast of Ireland.

Results and DiscussionFigure 2 shows the effect of crop types on thespecies richness (SR) and abundance of thecarabids, pollinators and plants. For the carabids,SR and abundance in the CT and OS were shownto be significantly higher than the other croptypes. A similar pattern was shown for pollinatorSR where the annual crops (OS and CT) weresignificantly higher than the perennial crop types(MG, MT, CG). For pollinator abundance, therewas only a significant difference found betweenOS and the other crop types. In contrast, for theplants, SR in CT was shown to be significantlylower than all other crops types. Clear differencesin SR and abundance emerge between the annualand perennial crop types where more resourcesavailable in the annual crops appear to favour bothcarabid and pollinator SR and abundances. On theother hand plant diversity and abundance isfavoured by stability of perennial crops. Cropeffects on the composition of the carabid,pollinator and plant communities are shown in theordinations outlined in Figure 3. All plantcommunities were significantly different from

each other (p<0.001) except those in MT and MG.For the carabids, no significant differences werefound between the communities of the OS andCT, and between MG, MT and CG, but significantdifferences were shown between the communitiesin the annual crops (OS and CT) and those of theperennial crops (MG, MT and CG) (p<0.001). Forthe pollinators, the OS communities were shownto be significantly different from the communitiesof all other crops (p<0.011), while no significantdifferences were found between the communitiesof all other crops. In general, Miscanthus wasshown to be most similar to other perennial crops,and most dissimilar to the annual tillage crops, i.e.the replacement of traditional crop types withMiscanthus did not result in novel communitiesexcept for the plants. Overall, this study showedthat growing Miscanthus did not have an obviousnegative impact on biodiversity as measured usingcarabids, pollinators and plants at field scale.However, it is important to note that replacingtillage crops with Miscanthus, thereby reducingoverall landscape heterogeneity, may result inbiodiversity losses at the landscape scale.

Crop type

CG CT MG MT OS

1015

20

a ab bb

CG CT MG MT OS

500

1000

1500

b da cc

CG CT MG MT OS

500

1000

1500

b da cc

CG CT MG MT OS

1015

2025

30

b aa aa

CG CT MG MT OS

1015

2025

30

b aa aa

CG CT MG MT OS

1.0

1.2

1.4

1.6

1.8

2.0

2.2

b ab bb

CG CT MG MT OS

1.0

1.2

1.4

1.6

1.8

2.0

2.2

b ab bb

CG CT MG MT OS

1015

20

a ab abb

CG CT MG MT OS

1015

20

a ab abb

CG CT MG MT OS

1.5

2.0

2.5

a ab bb

CG CT MG MT OS

1.5

2.0

2.5

a ab bb

Richness Abundance

Cara

bid

sP

ollin

ato

rsP

lan

ts

CG CT MG MT OS

1015

20

a ab bb

CG CT MG MT OS

500

1000

1500

b da cc

CG CT MG MT OS

500

1000

1500

b da cc

CG CT MG MT OS

1015

2025

30

b aa aa

CG CT MG MT OS

1015

2025

30

b aa aa

CG CT MG MT OS

1.0

1.2

1.4

1.6

1.8

2.0

2.2

b ab bb

CG CT MG MT OS

1.0

1.2

1.4

1.6

1.8

2.0

2.2

b ab bb

CG CT MG MT OS

1015

20

a ab abb

CG CT MG MT OS

1015

20

a ab abb

CG CT MG MT OS

1.5

2.0

2.5

a ab bb

CG CT MG MT OS

1.5

2.0

2.5

a ab bb

Richness Abundance

Cara

bid

sP

ollin

ato

rsP

lan

ts

CG CT MG MT OS CG CT MG MT OS

Figure 2. Boxplots highlighting the relationshipsbetween carabid, pollinator, and plant speciesrichness and abundance and crop type (CG, Ct,MG, MT, OS). Significantly different crop typesare shown by the letters (a, b, c, d) at the top ofeach graph.

a

b

c

Figure 3. MDS ordination of the carabid (a),pollinator (b), and plant (c) communities inrelation to crop type. Similarities based on BrayCurtis.

ConclusionsIn this study, the impact of growing Miscanthuson former grassland and tillage crops did notnegatively affect alpha-diversity. Future large-scale replacement of these crops with Miscanthusmay however affect gamma-diversity inagricultural landscapes. Maintaining land-usediversity in these landscapes must be a priority toensure climate change mitigation measures suchas bioenergy crops do not negatively impactfarmland biodiversity.

AcknowledgementsThis study is funded by the EPA as part of theSIMBIOSYS Project – www.simbiosys.ie.Special thanks to the participating farmers.

ReferencesClarke, KR, Gorley, RN, 2006. PRIMER v6: UserManual/Tutorial. PRIMER-E, Plymouth.Dauber, J., Jones, M.B., & Stout, J.C. (2010). Theimpact of biomass crop cultivation on temperatebiodiversity. Global Change Biology Bioenergy,2: 289-309.R Development Core Team (2008). R: A languageand environment for statistical computing.http://www.R-project.org.

How to measure the environmentaleffectiveness of the Rural EnvironmentalProtection Scheme?

J.A. FinnEnvironment Research Centre, Teagasc,Johnstown Castle, WexfordEmail: john.finn@teagasc.ie

IntroductionAgri-environment schemes in the EU are now oneof the most important policy mechanisms for theprotection of public goods, and offer payments tofarmers in return for undertaking managementpractices (measures) that are intended to maintain,enhance or restore the rural environment. TheRural Environment Protection Scheme (REPS)has become a widely adopted scheme (over54,000 participants in 2009), and provides animportant financial contribution to farm incomesin Ireland (e.g. DAFF, 2008; Connolly et al.,2009). Since 1994, REPS has paid a total of over€3.1 billion to Irish farmers, and paid over €330million in 2009. In return for these payments,participating farmers undertake a variety ofprescribed measures that are intended to benefitwater quality, soil quality, nutrient management,grassland management and biodiversity.

Since its inception in 1994, there has been strongdemand for evidence of the environmentaleffectiveness of the REPS. A number of differentforces are aligning that will likely result in variouspressures on agri-environment schemes. Theseinclude an increase in the number of EU MemberStates that will receive funding from the CommonAgricultural Policy and Rural DevelopmentProgramme, increased pressure on EU budgets,and increased pressure on the ability of individualmember States to provide co-financing. Inaddition, the EU Court of Auditors is due to reportits audit of the effectiveness of EU agri-environment schemes. The World TradeOrganisation also requires that the environmentalbenefits of agri-payments are clearlydemonstrated, to prove that such payments are notdisguised trade subsidies. At the same time, thereare strong suggestions of an increased provisionof public goods in the post-2013 CAP, and agri-environment schemes will be an important policyinstrument to achieve this (as well as others).

To date, there has not been a national-scale,comprehensive monitoring programme to measurethe environmental impacts of REPS. There will beincreasingly demanding requirements todemonstrate the environmental effectiveness (andespecially biodiversity benefits) of agri-

environment schemes. This desk study aimed tosupport decision-making about the appropriatedesign and implementation of an environmentalmonitoring programme for Irish agri-environmentschemes.

Materials and MethodsA desk study reviewed available publications thatare relevant to the environmental effectiveness ofREPS. The distribution of payments, across thedifferent environmental objectives of REPS1 andREPS 4, was examined to assess the extent towhich payments were allocated to specificenvironmental objectives (e.g. water, biodiversity,soil). A number of REPS options and measureswere selected as priorities for inclusion in anenvironmental assessment programme, and wereusually those with highest participation (as thesegenerally involved greatest expenditure). Broadaims were suggested for the sampling of each ofthe selected measures and options. Estimates ofthe number of field surveys and staff requirementswere used to estimate the total cost of anenvironmental monitoring programme.

Results and DiscussionThere has been a significant increase in therelative proportion of expenditure on basicmeasures for biodiversity-related objectivesbetween REPS 1 (~57%) and REPS 4 (~79%)(Finn, 2010; Finn and Ó hUallacháin, in press).This is not surprising given that most of themeasures associated with the original priorityobjective of REPS to protect water quality(largely through improved nutrient management)have since become part of the standardsassociated with cross-compliance, which are notpaid for. This clearly indicates that the majority ofREPS 4 payments is now associated withbiodiversity objectives. In addition,supplementary measures and options aredominated by biodiversity issues. Thus,measurement of the effectiveness of biodiversitymeasures and options should be a priority forenvironmental monitoring.

Note that Table 1 generally includes measureswith highest participation. It omits severalelements of REPS that: have insufficientparticipation to have an environmental effect;have insufficient existing information availablewith which to assess performance, or; require fieldexperiments rather than monitoring.

Overall, the monitoring of selected REPSmeasures, supplementary measures, biodiversityoptions and Measure A was estimated to costabout €3.4 million over a four-year period (Finn,2010). The monitoring programme would need to

recruit 18 different staff (eight of which would bepart-time).

For the whole programme, there would need to beat least 1500 different field surveys. Note thatthere is very different spatial distribution ofdifferent REPS measures, supplementarymeasures and options. Privileged access to theDepartment of Agriculture, Fisheries and Food(DAFF) REPS database and e-REPS would benecessary to help quickly identify the location andcontact details of randomly selected farms thatcontain specific measures or groups of measures.The effectiveness and cost of the monitoringprogramme would be very dependent on suchprivileged access. Privileged access to the REPSdatabase will be necessary for the design andimplementation of an effective and cost-efficientmonitoring programme. Without such access, thecosts would increase considerably.

ConclusionsThe average annual budget for the proposedmonitoring programme (~€0.86m) would be lessthan 0.25% of the recent annual expenditure onREPS (>€360m in 2009).There is considerable overlap and similaritybetween the existing REPS measures and options,and those included in the new Agri-EnvironmentOptions Scheme (AEOS) that will replace REPS.Thus, an assessment of the environmental impactsof REPS could be used to more quickly assess the

probable environmental effectiveness of similarmeasures that are implemented in the AEOS. Thecost of measuring the environmental performanceof REPS should be viewed as an investment insecuring the future of agri-environment schemesin Ireland. This would provide necessaryinformation to confirm the environmental benefitsof effective measures, and to implement anyrequired improvements to other measures. In thelong term, such an approach would ensure thatthese schemes and the participant farmers getappropriate credit for their successes.

AcknowledgementsThis project was funded by Teagasc RMIS 5757.

ReferencesConnolly, L., Kinsella, A., Quinlan, G. andMoran, B. (2009) National Farm Survey 2008.Carlow. Teagasc.

DAFF (2008) REPS Facts and Figures 2007.Dublin. Department of Agriculture, Fisheries andFood.

Finn, J.A. (2010) Monitoring the environmentalimpacts of the Rural Environmental ProtectionScheme: a scoping study. End of Project Report,RMIS 5757, Teagasc, Oak Park.

Finn, J.A. and Ó hUallacháin, D. A review ofevidence on the environmental impact of Ireland’sRural Environment Protection Scheme (REPS).Biology and Environment (in press).

Table 1. Suggested priorities for assessment in REPS 4 and indicative workload (no. of farmvisits) fieldwork associated with sampling of selected REPS measures and options. Estimatesare for environmental sampling only and assume independent visits to farms for eachmeasure or option (which could be reduced, see text for details).

REPS 4 measures and optionsNo. of Repsfarms

No. of non-REPS farms

Measure A (Natura 2000 and other priority sites) 250 0Measure 3 Watercourses and wells 90 30Measure 4: Farmland habitats 240 60SM2 Traditional Irish orchards - -SM5 LINNET 50 0SM8 Traditional grazing systems 70 40SM9 Clover swards 100 0SM10 Mixed grazing 100 252A Traditional hay meadow 50 202B Species-rich grassland 100 02E Control of invasive species 35 254A Creation of new habitat 70 04B Broadleaved tree planting 80 04C Nature corridor 70 04D Farm Woodland establishment 28 05C/5A New hedgerow establishment 70 405D Stone wall maintenance 60 359B Environmental management of setaside 60 30

Landscape aesthetics: assessing the generalpublics’ preferences for rural landscapes

P. HowleyRural Economy Development Programme,Teagasc, Athenry, GalwayEmail: peter.howley@teagasc.ie

IntroductionAt a broad level, the general public can becharacterised as nature friendly, that is, individualslargely acknowledge the intrinsic value of natureand its subsequent right to exist irrespective of itsfunctions for mankind. Individuals regard theirinteractions with what can be termed as naturallandscapes as more positive than their experienceswith landscapes that have been shaped to a largedegree by human interaction. This finding has beeninterpreted as supporting an evolutionary theory oflandscape preferences, whereby it is assumed thatsimilarities in responses to natural scenes outweighthe differences across cultures or smaller groups ofindividuals (Ulrich, 1993). There has, however,been widespread disagreement as to the validity ofthis consensus assumption. Specifically, muchresearch has found substantial individual and intergroup differences in landscape preferences (VanDen Berg et al., 1998). With this in mind, thecentral aim of this study was to gain greaterinsights into the individual characteristics thataffect preferences for a variety of landscapesettings.

Materials and MethodsA survey of 430 individuals living in Ireland wasconducted in the summer of 2010. The respondentswere asked to indicate their preferences for rurallandscapes by rating 47 landscape images on ascale from 1 (not very highly) to 6 (highly). Thephotographs themselves were selected with the aimof representing a broad geographic and thematicrepresentation of rural landscapes in Ireland.Environmental value orientations were measuredby including a series of attitudinal statementsconveying two pro-environmental values -ecocentrism, anthropocentrism - andenvironmental apathy. The statements relating toecocentrism were devised to capture theimportance of nature and the amenity as well asvisual value of the landscape whereas theanthropocentric statements refer to a morefunctional view of the landscape - one thatemphasises the importance of using the landscapefor agricultural production.

Data analysisA factor analysis (principal component analysiswith varimax rotation) was employed on theattitudinal statements designed to capture

environmental value orientations and also onrespondents mean scores of the landscape images.As expected, the factor analysis of the attitudinalstatements resulted in three factors with aneigenvalue > 1, together explaining 61 percent ofthe variance. The statements relating toecocentrism loaded highly on the first factor and assuch this factor was termed ‘ecocentric value’. Thestatements relating to environmental apathy loadedhighly on the second factor and finally thestatements relating to anthropocentrism loadedhighly on the third factor. Therefore theseindividual factors were labelled as ‘environmentalapathy’ and ‘anthropocentic value’ respectively.These individual factor variables were utlised in anOLS regression model in order to examine theirrelative influence on preferences towards a varietyof landscape types.

A factor analysis was also employed onrespondents mean ratings of the 47 landscapeimages. A five factor solution proved to give thebest solution. The derived factors representindividuals’ distinct preferences towards differentfeatures of the landscape. Based on an analysis ofthe factor loadings, the factor variables werelabelled as ‘intensive agriculture’, ‘extensiveagriculture’, ‘wild nature scenes’, ‘culturallandscapes’ and ‘water related landscapes’.Individual factor scores for each category oflandscape were used as a dependent variable inseparate regression models designed to examine ifthey were any socially differentiating factorsaffecting individual landscape preferences. Factorscores representing respondents’ differentenvironmental value orientations were alsoultilised as explanatory variables in the followinganalysis to examine if these along with backgroundsocio-demographic factors influenced the generalpublics’ landscape preferences.

Results and DiscussionWater-related landscapes attracted the highestmean scores by respondents. Cultural-relatedlandscapes are also highly regarded by respondentsas all of the images in this category also attractedrelatively high mean scores. In relation to theagricultural landscapes, respondents rated all ofthese quite highly as all the mean scores were atthe upper end of the six-point scale. Theagricultural landscapes that respondents appearedto like least, however, were the more intensivefarming landscapes. This supports findings in avariety of other studies which suggest that modernintensive farming landscapes are less attractive tothe general public due mainly to the homogeneityof this type of landscape.

Multivariate regression analysis was used toexamine what factors influenced respondents’preferences for each of the landscape types derivedfrom the factor analysis. The dependent variablewas individuals’ factor scores for each of thederived 5 perceptual categories of landscape. Theresults from each of the 5 regression models arepresented in table 1. Age was the socioeconomicvariable that was perhaps the strongest predictor ofpreferences in that it was statistically significant indetermining preferences for three of the landscapecategories (intensive and extensive farming andwater related landscapes). The positive relationshipbetween age and both agricultural landscapes couldbe reflective of generational differences in cultureand upbringing with relatively elderly respondentsmore likely to be familiar with agriculturallandscapes. Age had a negative association withwater-related landscapes and this could beattributable to older people’s greater vulnerabilityto the dangers of this type of landscape.

Environmental value orientations were perhaps themost significant determinant of landscapepreferences as these were found to strongly affectpreferences for each of the landscape typesexamined. Environmental value orientations aredefined as individual or societal beliefs about theimportance of the natural environment and inparticular how the natural world should be viewedand treated by humans (Reser andBentrupperbaumer, 2005). An ecocentric valuewas found to have a positive impact on all thelandscape types examined (extensive farminglandscapes, cultural landscapes, wild nature scenesand water related landscapes) with the exception ofintensive farm landscapes where it was not foundto have a statistically significant impact. Theselandscape types may be preferred over intensivefarming landscapes by respondents with a strongecocentric value because of their strong amenity,ecosystem or wildlife aspects. An anthropcentricvalue orientation and environmental apathy werefound to have a negative effect on preferences for‘wild nature scenes’. It could be that the relativelyunproductive nature of this type of landscapemakes it unattractive for respondents with either ofthese types of value orientations. Finally,environmental apathy was also related topreferences for intensive farming landscapes asrespondents who were indifferent to environmentalissues were more likely to rate this type oflandscape in terms of beauty highly

Table 1. OLS regression model examining factorsinfluencing landscape preferences (statisticallysignificant variables highlighted in bold).

Coefficient Intensive Extensive

Wildnaturescenes Cultural

Waterrelated

landscapes’Age 0.03* 0.04*** 0.01 0.002 -0.04**

Females 0.08 0.37*** -0.13 -0.12 0.07Social class 0.11 -0.19** 0.27*** 0.02 0.06Rural 0.30** 0.31*** 0.08 -0.09 0.21*

Town 0.35*** -0.03 -0.09 0.05 0.07Farmingbackground 0.13 0.35*** -0.10 -0.11 -0.12Ecocentricvalue 0.03 0.26*** 0.21*** 0.20*** 0.13***Anthropocentric value -0.04 0.10** 0.13*** 0.03 0.020Environmental apathy 0.09* 0.06 0.19*** -0.19*** -0.20***

* significant at 10 percent level, **significant at 5 percent level, ***significant at 1 percent level

ConclusionsThe landscape can be viewed as an economicresource and as a local public good in that itprovides amenities and supports recreational aswell as productive activities. Policy perceptions ofrural landscapes have changed over time from sitesof mass agricultural commodity production toareas of socio-cultural, economic and ecologicaldiversity in which a range of goods are bothproduced and consumed. Land use policy can beimproved if decision makers in both theenvironmental and agricultural sectors are betterinformed about the landscape preferences andattitudes toward the environment among varioususer groups. The results presented here suggest thatthere are distinct differences in terms of landscapepreferences between different demographicgroupings and also depending on individuals’environmental value orientations. Accordingly, instudying landscape preferences in particular areasit will be necessary to consider the personalcharacteristics of the population as well as thephysical aspects of the landscape. Moreover interms of land use policy, given the diversity ofpreferences a one-size-fits-all approach will notmeet the general publics’ needs and desires.

ReferencesUlrich, R.S. (1993) Biophilia, biophobia, andnatural landscapes. In: Kellert, S.R., Wilson, E.O.(Eds.), The Biophilia Hypothesis. Island Press,Washington, DC, pp. 73–137.

Van Den Berg, A.E., Vlek, C.A.J. and Coeterier,J.F. (1998) Group differences in the aestheticevaluation of nature development plans: amultilevel approach. J Environ Psychol. 18: 141–157.

Reser, J.P. and Bentrupperbaumer, J.M. (2005)What and where are environmental values?Assessing the impacts of current diversity of use of‘environmental’ and ‘World Heritage’ values. JEnviron Psychol. 25: 125–146.

Cessation of grazing - the effects on land snailcommunities in farmed grassland, scrub andwoodland habitats in the Burren region in thewest of Ireland

M.P. Long1, E.A. Moorkens2 and D.L. Kelly1

1Botany Department, Trinity College, Dublin 2,253 Charleville Square, Rathfarnham, Dublin 14Email: longma@tcd.ie

IntroductionThe Burren is famous for its flora and fauna. Itsimpressive biodiversity is indebted in no small wayto the agricultural traditions of the area (Dunford,2002). The limestone pavements and the species-rich grasslands are among the most remarkableaspects of the Burren. These habitats are now set ina changing landscape. Many areas are being takenover by encroaching hazel scrub which threatenssome aspects of the biodiversity of the Burren (TheHeritage Council, 2006). The scrub also interfereswith farming by blocking trackways used by stockand taking over valuable farmland.

One of the main theories about why this ishappening is because of changes in farming. Acombination of factors such as the use of less-hardy animal breeds, farmers often needing towork off-farm and the changeover from beef cattleto suckler cows has resulted in a general decreasein the grazing pressure on some of the mostvaluable Burren habitats (Dunford and Feehan,2001; Dunford, 2002; Williams et al., 2009).

The current study investigates the effects that thecessation of grazing might have on biodiversity,focussing on vascular plants and land snails. Veryfew studies have used grazing exclosures in orderto investigate the effects of changes in landmanagement practices on molluscan communities.Of two such studies, higher densities of snails werefound in ungrazed areas of boreal forests inFennoscandia, compared to plots grazed by mooseand/or reindeer (Suominen, 1999), and negativeimpacts of cattle grazing on mollusc densities werefound in fens in the UK (Ausden et al., 2005).

Materials and MethodsA network of twelve fenced exclosures (each 20 x20m) was set up across the Burren region in 2006.The fences prevent access by large grazers (mainlycattle and goats) and were placed in three types ofhabitat: rough grasslands, areas with low orscattered hazel scrub, and hazel woodlands. Besideeach fenced area is an unfenced plot of similar sizewhich acts as a control. Plants and snails (and otheraspects of biodiversity) have been monitored sincethe set-up in both the fenced and control plots.

Molluscs were sampled in both 2006 and 2008using 25cm x 25cm quadrats, five in each plot. Alllow-growing vegetation, litter (dead plant material)and loose soil were removed, dried and sieved intodifferent size fractions. The molluscs wereremoved and identified. Slugs were not included inthis study.

Results and DiscussionSummary of vegetation findingsThe strongest changes recorded in the vegetationcame from the grassland sites, where there was asignificant decrease in both plant richness anddiversity, and a significant increase in the amountof litter. The cessation of grazing appears to havehad a negative effect on grassland plantcommunities, at least in the short term. Thewoodlands showed an increase in diversity. Resultsare more complex for the scrub habitat, and longer-term study is needed.

Findings on snail diversityNo mollusc species were lost or gained betweenthe two sampling periods, although many speciesexhibited changes in abundance. The mean numberof snails collected per quadrat increased by almost50% inside the fenced plots between the first andsecond sampling periods (i.e. between 2006 and2008), and there was only a very small change inthe control plot numbers (a decrease of 3%).

The largest and most consistent changes were seenin the grassland sites, with significant increases inthe numbers of snails collected across all quadrats(p=0.03, one-tailed paired t-test) (Fig. 1). Thespecies which showed the greatest change wasColumella aspera. It appeared for the first time inthree of the grasslands sites in 2008. This is aspecies of ‘rough’ habitats generally, such asuncultivated grasslands (Kerney, 1999). Thegeneral trend was for an increase in molluscnumbers and species across most grassland sites.

Multivariate analysis using NMS (non-metricmultidimensional scaling) showed a definite shiftin the species composition of the mollusccommunities over the two-year period. In theresulting ordination, quadrats sampled in 2006tended to separate from those sampled in 2008(Fig. 2). The main factors associated with this shiftwere found to be cover of litter and vegetationheight, both of which increased substantially in theabsence of grazing.

0

1

2

3

4

5

6

7

8

9

Fenced plot Control plot

Incr

ease

inno

.ind

ivs

Figure 1. Average changes in numbers ofindividuals recorded in the grasslands (±s.e.).

Veg hgtBare earth

Litter

No. indivs

Axis 1 (30%)

Axis

2(2

3%)

Figure 2. NMS ordination of mollusc data fromfenced grassland plots. Each point corresponds to aquadrat. Figures in brackets on axis labels are thepercentage of the variation in the distance matrixwhich is explained by this axis. The mostinfluential variables are overlaid. Open circles,2006; closed triangles, 2008.

ConclusionsWhen grazing was removed experimentally fromgrassland habitats in the Burren, the snails seemedto benefit from the longer vegetation and denserlitter which resulted. It is likely that the litterprovides extra food, shelter and moisture for thesnails, and thus conditions have improved (at leastfor certain species) within the fenced exclosures atthe grassland sites.

These findings contrast with the decreases invascular plant richness and diversity which wereseen in the grasslands, and thus they highlight theimportance of assessing a suite of taxa wheninvestigating the effects of changes in managementpractices on biodiversity. If we compare these

findings (from grasslands) with those of Ausden etal. (2005) (from fens), we can see that the effectsof the removal of grazing on mollusc communitiesmay also be habitat-specific. This again points tothe importance of broad-scale studies, particularlywhen landscape-scale changes are in question.

The exclosures set up during this study provide avaluable tool for monitoring long-term vegetationand landscape change in the Burren into futuredecades. It is hoped that this work will becontinued into the longer term.

AcknowledgmentsWe are very grateful to all the landowners involvedin this project. Funding came largely from the EPA(BioChange project, www.biochange.ie), withadditional support from the National Parks andWildlife Service and SYNTHESYS, the EuropeanUnion-funded Integrated Activities grant.

ReferencesAusden, M., Hall, M., Pearson, P. &Strudwick, T.(2005). The effects of cattle grazing on tall-herbfen vegetation and molluscs. BiologicalConservation 122: 137-326.

Dunford, B. (2002). Farming and the Burren.Teagasc, Ireland.

Dunford, B. &Feehan, J. (2001). Agriculturalpractices and natural heritage: a case study of theBurren Uplands, Co. Clare. Tearmann: Irishjournal of agri-environmental research 1: 19-34.

Kerney, M. P. (1999). Atlas of the Land andFreshwater Molluscs of Britain and Ireland.Harley Books, Essex. England.

Suominen, O. (1999). Impact of cervid browsingand grazing on the terrestrial gastropod fauna inthe boreal forests of Fennoscandia. Ecography 22:651-658.

The Heritage Council (2006). Assessment ofLandscape change and effects on Archaeology andan Assessment of Habitat Survey in the Burren, Co.Clare. The Heritage Council, Kilkenny, Ireland.

Williams, B., Parr, S., Moran, J., Dunford, B. &Ó'Conchúir, R. (2009). The Burren - farming forthe future of the fertile rock. British Wildlife 21: 1-9.

Development and application of the Agri-environmental Footprint Index (AFI)

G. Louwagie1, J.A. Finn2, G. Purvis1 and G.Northey2

1School of Agriculture, Food Science & VeterinaryMedicine, University College Dublin, Ireland2Teagasc, Environment Research CentreJohnstown Castle, Wexford, IrelandEmail: geertrui.louwagie@gmail.com

IntroductionThe Common Monitoring and EvaluationFramework (CMEF) developed by the EuropeanCommission provides a unified framework formonitoring and evaluation of all rural developmentinterventions for the programming period 2007-2013. Given the many customised implementationsof the EU rural development instrument (RuralDevelopment Regulation (EC) 1698/2005,amended by Council Regulation (EC) 74/2009) inthe interests of subsidiarity and effectiveness, it isenvisaged that the CMEF will need to besupplemented, as specific contextual factors indifferent regions call for locally-adapted, context-specific indicators and evaluation (e.g. due tovariable geology and soils, topography, climate,farming type and tradition). The Agri-environmental Footprint Index (AFI) wasdeveloped as the primary output of an FP6-fundedproject, to provide exactly this flexibility, whilstusing a common conceptual evaluation structureand harmonised process (see Purvis et al., 2009).

Materials and MethodsThe methodology developed in the AE-FOOTPRINT project involves the construction ofan AFI, allowing the combination of variousindicators reflecting the environmentalperformance of a particular farm. The approachemploys components of multi-criteria analysistechniques to provide a means of combiningindicators corresponding to a variety of farmmanagement activities and relating to a range ofenvironmental objectives. The methodologyincorporates the participation of stakeholders andtechnical advisors in designing a customised formof the AFI relevant for each particular application.In the methodology, stakeholders validate theassessment criteria and provide a series of weightsallowing combination of different components ofenvironmental performance. Such input of specific,technical and local knowledge ensures theevaluation is appropriate to the local agri-environmental context. The participatory process isdescribed at length by Mortimer et al. (2009) in anAFI User’s Manual.

The AFI also relies on a matrix structurecomprised of three major environmental issues(Fig. 1) crossed with three management strategiesor management domains (crop and animalhusbandry, physical farm infrastructure, andnatural and cultural heritage). Through aparticipatory process, this universal matrixstructure can be customised to any agri-environmental context. The resulting nine-dimensional matrix accommodates two essentialbuilding blocks of the AFI: an assessment criteriamatrix (ACM) and an indicator matrix (IM). TheACM gives a context-specific description of theassessment to be made; the IM defines the meansto make the assessment. (For details, see Purvis etal., 2009; Mortimer et al., 2009).

AFI applications in IrelandOne of the case study areas was in Ireland, wherethe AFI methodology was tested by evaluating theeffectiveness of a broad, multi-objective agri-environmental scheme (AES), the Irish RuralEnvironment Protection Scheme (REPS). Wereport two contrasting farming situations:relatively intensive dairy farming in County Cork,and comparatively extensive dry-stock(beef/sheep) farming in County Sligo. Thecustomised AFI was created through an interactiveprocess with stakeholders, and was similar for bothapplications. Stakeholders consisted of a group ofenvironmental, agricultural, natural/culturalheritage, production and agri-environmental policyspecialists, along with agricultural advisors andfarmers. Assessment criteria were largely definedbased on the REPS objectives and associatedobligations, as REPS provided the evaluationbackground for the presented applications.

This scheme is very holistic and addressed all ninedimensions of the AFI matrix structure. Indicatorswere selected based on the agreed assessmentcriteria. Transformation functions were developedto map all indicator values onto a common scoringscale (0 to 10). However, as some ACMdimensions contained up to 14 criteria, it wasnecessary to devise more complex multi-metricindicators in order to evaluate all identifiedconcerns, whilst still meeting the desired target ofusing no more than a maximum of five to sixindicators per matrix dimension. This allowed ‘realworld’ interactions and relationships betweenmultiple agri-environmental concerns and farmingpractices to be aggregated into a single indicatorscore. Technical specialists were regularly askedfor their feedback to secure a sensible and coherentstructure for the transformation functions.Stakeholders weighted issues, managementstrategies and the underlying aspects of indicatorsindividually. These weights are perceptions of

relative importance when policy priorities are notor not sufficiently defined. Where possible, aconsensus of final issue and management strategyweights was reached among the stakeholders; inother cases the mean of the individual weights waschosen. The AFI was quantified by multiplyingscores with weights (weighted sum), subsequentlyat the three levels of indicators, managementstrategies and issues. In both contexts, the AFI wascalculated for a small sample of farms of whichhalf had a REPS agreement. As a proof-of-conceptapplication, data were collected for indicatorsrelevant to the ACM for REPS and non-REPSfarms in Sligo and Cork. The environmentalcriteria that were used went beyond those based onREPS, to measure wider environmental impacts ofthe scheme.

Results and DiscussionIn Cork, data were collected from eight REPSdairy farms and eight non-REPS dairy farms. Inthe Sligo region, indicator data were collected onten REPS dry-stock farms and ten non-REPS dry-stock farms. In the application of the AFI in Cork,the mean AFI scores of the REPS farms (4.72) wassignificantly greater than the mean AFI score(3.78) of the non-REPS farms (Fig. 1). In Sligo,the mean AFI score of the REPS farms (5.74) wassignificantly (p=0.05) higher than that of the non-REPS farms (5.00). At the level of individualissues, the REPS farms in the sample consistentlyoutperformed the non-REPS farms. Note that allthe farms in the Cork case study area scoredsubstantially lower than the livestock farms in theSligo case study area. This interpretation requiressome care, due to the low sample sizes, and due tothe two slightly different forms of the AFI used(weighting and a small number of indicatorsdiffered).

0

1

2

3

4

5

6

7

1 2 3 40

1

2

3

4

5

6

7a. Cork b. Sligo

natural resources

biodiversity

landscape

REPS non-REPS REPS non-REPS

0

1

2

3

4

5

6

7

1 2 3 40

1

2

3

4

5

6

7a. Cork b. Sligo

natural resources

biodiversity

landscape

REPS non-REPS REPS non-REPS

Fig. 1. Comparison of AFI scores (illustratingperformance among three different environmentalissues) in participating REPS farms and non-participating farms in the Cork and Sligo regionsof Ireland

Overall, these results have to be treated withcaution, as the sample size of up to 20 farms isvery limited, and is not representative of the

national-scale application of REPS (which hasabout 50,000 participants). The AFI approach wasapplied in six other EU member states, and theseresults are representative of the outcomes.

ConclusionsREPS farms attained higher AFI scores than non-REPS farms in both applications of themethodology in Ireland. Note that the sample sizein this study was too small to be a nationallyrepresentative comparison of REPS and non-REPSfarms. There may also be biased participation offarms with higher levels of environmental quality.Nevertheless, this work demonstrates thefeasibility of constructing an AFI for theevaluation of an AES with multiple environmentalobjectives. The weighting process also permitsprioritisation of environmental objectives that maydiffer among farming systems, among regions, andamong different AESs.Overall, the AFI methodology developed by theAE-Footprint project has a number of features thathave the potential to contribute to the aims of agri-environmental evaluation. The method also offersthe advantage of being quantitative andtransparent. Importantly, the participatoryapproach can also help increase the quality ofpolicy dialogue among policymakers, stakeholdersand farmers.

AcknowledgmentsWe thank the many technical specialists andstakeholders who made invaluable contributions tothis study. In particular, we acknowledge theextensive help of Mr Ben Wilkinson and Mr DenisCarr in collecting farm data. We thank the farmerswho generously agreed to participate in this study,and gave us permission to survey their farms.This study presents results obtained within the EUFP6 project SSPE-CT-2005-006491, The Agri-environmental Footprint: Development of acommon generic methodology for evaluating theeffectiveness of European Agri-environmentalSchemes (AE-Footprint). The views expressed arepurely those of the authors and may not in anycircumstances be regarded as stating an officialposition of the European Commission.

ReferencesMortimer et al. (2009) The Agri-environmentalFootprint Index: User’s Manual. URL:http://www.footprint.rdg.ac.uk/afimanual/.

Purvis, G., Louwagie, G., Northey, G., Mortimer,S., Park, J., Mauchline, A., Finn, J.A. et al. (2009)Conceptual development of a harmonised methodfor tracking change and evaluating policy in theagri-environment: The Agri-environmentalFootprint Index. Environ. Sci. & Policy 12, 321-337.

Can current Irish riparian managementpractices enhance ground beetle diversity?

D. Madden,1, S. Harrison1, J.A. Finn.2, D. ÓhUallacháin2

1Department of Zoology, Ecology and PlantScience, University College Cork, 2Teagasc,Johnstown CastleEmail: madden_d@yahoo.com

IntroductionRiparian zones have high levels of biodiversity andplay an important role in the abundance, diversityand functioning of floral and faunal communitiesin agri-ecosystems. The nature of riparianvegetation in agricultural settings is largely afunction of management practices. The RuralEnvironment Protection Scheme (REPS) in Irelandincludes measures to exclude grazing stock fromriparian areas through fencing (1.5m from the edgeof the stream), but prescriptions for riparianmanagement within these fenced zones are varied.Farmers may plant trees within the fenced area, butthey can also cut back vegetation within the fencedarea. The impact of this management onbiodiversity is uncertain, given the lack of specificbiodiversity aims, representative target species orstandardised monitoring system to identify anyimpacts.Carabid ground beetles are widely considered to besuitable indicators of conditions in riparian habitatsdue to the sensitive and quick reaction of manyspecies to environmental change. Speciesassemblages can reflect differences in both fluvialhydrology and micro-habitat conditions.The aims of this study were to assess the impactsof vegetation type on riparian carabid beetlecommunities in grassland agricultural catchments.The results will inform policy makers on therelative success for biodiversity enhancement ofthe current riparian management guidelines.

Materials and MethodsStudy sitesTen farms in SE Ireland, managed under the REPSscheme, were selected for sampling. Each farm hada 1st or 2nd order stream within its boundaries,along which were present three riparian vegetationtypes. These were classified as grass (lowherbaceous vegetation), scrub (vegetationdominated by bramble and/or gorse) and woodland(dominated by mature trees such as alder andwillow). Each stream within all 10 farms possessedall three vegetation types.

Carabid and Botanical samplingPitfall traps were used to sample carabids withineach of the three vegetation types along eachstream, during 2-week periods in both June and

September, 2008. Seven pitfalls traps were placedparallel to each stream, within 1m of the bank edgeand with 2m between each trap. Beetle data foreach site was combined from both collectionperiods. Ground vegetation for each site wassampled with ten 0.25m2 quadrats, randomlylocated parallel to (and within 1.5m of) eachwatercourse bank edge. Upper-storey vegetationwas given Braun Blanquet values for the entire20m plot.

Data analysisCarabid abundances were standardised to trapmean per site to correct for lost samples. Speciesrichness (SR) was calculated using rarefaction(ECOSIM software, Gotelli & Entsminger 2001),which eliminated variation in SR due todifferences in sample size and sampling effort.This method repeatedly resamples a pool ofsamples at random, to produce both abundancecurves (abundance of individuals per trap) anddiversity curves (SR per trap). Rarefaction curvesplotted ‘expected’ SR across all levels of sampledabundance and number of samples, allowing for astandardized comparison of richness amonghabitats. Analysis of similarity, ANOSIM, wasused to determine if there were significantdifferences in assemblages between habitat typesfor carabid and floral communities. ANOSIMproduces an R-value statistic, which if approaching1 indicates strongly distinct assemblages, and ifclose to zero indicates that the assemblages arevery similar. Two-way ANOVA or non-parametricFriedman test with Tukey post hoc tests were usedto examine differences in SR and abundance.

Results and DiscussionA total of 1943 beetles, from 51 species(approximately 24% of the Irish carabid fauna),were identified. The most abundant species werePterostichus melanarius (38%) and P. madidus(17%), both habitat generalists. Only three specieswere riparian specialists, all occurring on 50% orfewer of the study farms (Table 1). Furthermore,no habitat appeared to be more ‘riparian’ incharacter than any other (Table 1). The streamsidehabitats in our study clearly did not represent trueriparian habitats for carabid beetles. This may bedue to the highly modified nature of the streamsidehabitats, their very small spatial scale, or theirisolated and fragmented nature. Anderson (1996)has suggested that the riverine riparian carabidfauna in Ireland are disjointed and scattered due toarterial drainage schemes across large areas.Although 19 carabid species were confined to asingle habitat type, 14 of these were singletons, 16were found on only one farm and none showed anassociation with a habitat after analysis withIndicator Species Analysis, indicating that this may

be a site effect, rather than a more general habitateffect. Other studies have shown that differentriparian habitats can contain very distinctcommunities (Cole et. al., 2008; Plachter andReich, 1998; Eyre et al., 2001). The low betaspecies diversity in our study reflects thedominance by eurytopic species which may be dueto the highly modified nature of the streamsidehabitats in these agricultural catchments.

Table 1. Abundance and frequency of occurrenceof carabid beetles that are riparian specialistsamong farms and habitat types. (habitats: G, grass;S, scrub, W, woodland).

SpeciesTotalNumber ofIndividuals

TotalNumberofFarms

Abundancein eachhabitatG,S,W

% ofTotalCarabidCatch

Paranchusalbipes

14 5 2, 3, 9 0.7

Pterostichusnigrita

8 2 2, 1, 1 0.4

Trechoblemusmicros

8 5 3, 1, 2 0.4

Among habitats, carabid abundance was highest ingrassland and richness highest in scrubland (Table2). Grassy field margins have been found to haveboth low (Cole et al., 2008) and high densities ofcarabids (Kromp and Steinberger, 1992) whencompared to infield habitats. Our results indicatethat the managed grassy habitats within fencedriparian zones, although not as speciose as otherhabitats, may provide highly suitable habitat forgeneralist agricultural species. When sites from allfarms within each vegetation type were combined,scrub cover provided the highest expected speciesrichness per number of individuals, whereas grassand wood were not significantly different fromeach other e.g. 400 individuals have an expectedSR of 36 in scrub habitats, 25 in grass habitats, and24 in wood habitats.

Table 2 Impact of vegetation type on carabidabundance (mean ± SD per trap) and speciesrichness (plot mean after rarefaction by sample).Values sharing letters are not significantly different.

Grass Scrub Wood ANOVAAbundance 7.1 ±

53.8 ± 2

5.1 ±3.3

P=0.063

Speciesrichness

5.5 ±1.9ab

6.6 ±1.6a

4.7 ±0.9b

P<0.05

The ANOSIM analysis confirmed the markedfloristic differences between grassland, scrub andwoodland riparian habitats (R-value of 0.86).However, this was not reflected in the carabidassemblages, which showed very little differencebetween habitats (R-value of 0.119). The threevegetation types along study streams, although

floristically distinct, clearly did not provide fordifferentiation among carabid beetle communities.

ConclusionsThe managed streamside habitats in our studycontained very few riparian carabid specialists.There was also little difference in the carabidcommunities between the three vegetation types,despite distinct floristic differences. Together,these results likely reflect the highly modifiednature of streamside habitats in many agriculturalcatchments.The management practice of fencing agriculturalstreams at this scale is unlikely to significantlyenhance riparian carabid diversity at the farm, orthe regional scale. Removing the disturbance ofcattle by fencing is unlikely to enhance thenumbers or diversity of riparian specialists,because the colonisation of riparian habitats byspecialists is hindered by the low carabid diversityin adjacent areas. Although fencing (with nosubsequent management) will likely lead tostreamsides being dominated by woody andscrubby vegetation, rather than grassy (grazed)vegetation, it seems unlikely that this would causea significant reduction in the number of grasslandspecialist species.

AcknowledgmentsThis project was funded by the Department ofAgriculture, Fisheries and Food under the NationalDevelopment Plan 2006 Research Stimulus Fund.

ReferencesAnderson, R. (1996). Species Inventory forNorthern Ireland: Carabid Beetles. Research andDevelopment Series, EHS.Brose, U. (2003). Bottom-up control of carabidbeetle communities in early successional wetlands:mediated by vegetation structure or plant diversity?Oecologia 135: 407-413.Cole et al. (2008). The influence of riparian bufferstrips on carabid beetle (Coleoptera, Carabidae)assemblage structure and diversity in intensivelymanaged grassland fields. Biod. & Cons. 17(9):2233-2245.Eyre, M.D., Luff, M.L. & Phillips, D.A. (2001).The ground beetles (Coleoptera: Carabidae) ofexposed riverine sediments in Scotland andnorthern England. Biod. & Cons. 10, 403-426.Kromp, B. & Steinberger, K. H. (1992). Grassyfield margins and arthropod diversity: a case studyon ground beetles and spiders in eastern Austria(Coleoptera: Carabidae; Arachnida: Aranei,Opiliones). Agr., Ecosys. & Env. 40: 71-93.Plachter, H. & Reich, M. (1998). The significanceof disturbance for populations and ecosystems innatural floodplains. Proc. Int. Symp. RiverRestoration, Tokyo, pp. 29-38.

Do cattle drinking points contribute tostream pollution levels in Ireland?

D. Madden1, D. Ó hUallacháin2, J.A. Finn2 and S.Harrison1

1Department of Zoology, Ecology and PlantScience, University College Cork, 2EnvironmentResearch Centre, Teagasc, Johnstown CastleEmail: madden_d@yahoo.com

IntroductionThe Water Framework Directive (WFD)(2000/60/EEC (Council of the European Union2000)) requires that water quality be managed andmonitored within river basins by a coordinatedeffort inclusive of all public authorities. TheDirective aims to achieve at least a ‘good status’ inall waters by 2015. In Ireland, the RuralEnvironment Protection Scheme (REPS) has beenthe main agri-environment scheme over the past 17years (Department of Agriculture, Food andForestry 2007) and so can be considered animportant vehicle for the protection ofwaterbodies. Within REPS and the Agri-environmental Options Scheme (AEOS, the currentreplacement scheme for REPS) farmers arerequired to prevent unrestricted cattle access towatercourses, but they are allowed to provide asingle drinking access point within each field.There are many additional water protectionmeasures required when participating in REPS,such as effective nutrient management, all ofwhich can contribute to reducing the pollution loadto surface waters. Anthropogenic effects onreceiving waters can range from the catchmentscale (e.g. intensive catchment agriculture) to thestretch scale (e.g. local livestock access). Atpresent however, there is little published data onthe effects of cattle drinking access points(CDAPs) as pollution pathways and sources toheadwater streams.

The aim of this study was to quantify the extent towhich CDAPs impact water quality in agriculturalheadwater streams. We investigatedmacroinvertebrate communities and waterchemistry parameters both upstream anddownstream from CDAPs on streams runningthrough grassland farms in south-east Ireland. Wesampled macroinvertebrates, because they arewidely used as indicators of stream and riverhealth, are considered good indicators of waterquality, and play an important role in the food webof streams and rivers. They are relativelyimmobile, easy to sample, relatively long lived,respond to a wide variety of pollutants, and canreflect long and short term trends (Chadd &Extence, 2004).

Materials and MethodsStudy sitesForty streams in the southeast of Ireland wereselected for sampling. The average annual rainfalltotal in the southeast has been lower than the restof the country over the past 50 years and farmingpractices are relatively intensive when compared towestern regions. The study was carried out on first-and second-order streams. These streams, whichare expected to be more impacted by intermediatepollution sources such as CDAPs, typicallyaccount for >95% of the total processing andtransportation of materials (Maitland, 1979). Themajority of streams were located on individualfarms, and sampling points were over 500m apart.

Water chemistry and macroinvertebrate samplingSampling was carried out from July-September2009. At each site, area of bare ground, adjacentland-use variables, stream physical attributes, andriparian vegetation was recorded. Water chemistry(total phosphorus, total nitrogen, reactivephosphorus, ammonium, nitrate, dissolved oxygen,conductivity and temperature), sedimentcharacteristics and macroinvertebrate communitieswere sampled at a single site upstream anddownstream from the drinking points at locationswhere stream flow and riparian vegetation weresimilar. Macroinvertebrates were sampled at thenext riffle point downstream of the CDAP, usuallyat a distance of 5-7 times the width of the stream,and immediately upstream where substrate andriparian habitat were most similar to thedownstream sampling location. Macroinvertebrateswere sampled using a 5-minute kick-samplingtechnique, covering approximately 3m2 of rifflehabitat.

Data analysisThe Small Stream Risk Score (SSRS) and Q-values are commonly used biotic indices of waterquality in Ireland and each was assigned tomacroinvertebrate communities at all samplinglocations. Up- and down-stream samples werecompared using t-tests of the water andmacroinvertebrate data. Because multiple t-testswere conducted, all comparisons of differencesbetween upstream and downstream data sets wereonly accepted as significant if p < 0.01.

Results and DiscussionForty CDAPs were sampled with a meanindividual area of bare or poached ground of14.6m2 (±10.9 SD), ranging from 0 to 50m2. Therewere no significant differences between upstreamand downstream in any water chemistry parameter,sediment or macroinvertebrate parameter (Table1). Longer-term continual sampling of waterchemistry or sampling over the duration of a high

rainfall event may be needed to detect any inputsfrom CDAPs.

Table 1. Mean (± s.e.) water nutrient parameters(mg/l), both upstream and downstream of CDAPs.No parameters showed a significant effect, p<0.01.

Upstream Downstream

N-NH4 0.04 ± 0.008 0.04 ± 0.01

N-NO2 0.01 ± 0.003 0.01 ± 0.003

RP 0.02 ± 0.004 0.02 ± 0.005

TN 5 ± 0.55 5 ± 0.54

TON 4.2 ± 0.53 4.1 ± 0.53

TP 0.04 ± 0.007 0.05 ± 0.01

The majority of sampled streams returned a Q-value below 4, indicating poor water quality. 10%of streams produced good water quality scores(Q4) (Fig. 1). There were few clear differencesbetween sites that were upstream of drinkingpoints, and those that were downstream. This wasalso the case for those streams with higher waterquality (Q4).

0

10

20

30

40

50

60

70

at risk or Q3 probably at riskor Q3-4

probably not atrisk or Q4

Nu

mb

er

of

sa

mp

les

SSRS

Q-value

Fig. 1. (A) Number of samples (n=80) assigned toeach Q-value and SSRS category. The broken lineindicates a boundary between high and lowerquality streams.

As highlighted by Q-values, SSRS scores showedthat water quality was generally low in themajority of the sampled streams. Again, only 10%of streams were within the ‘probably not at risk’category, with two streams showing a decrease inwater quality downstream and two streamsshowing an increase in water quality downstream.

ConclusionsThere was no significant impact of CDAPs onstream water quality and biological communities inthis sample of study streams. However, a majorityof these small streams in our sample were ofrelatively low water quality, irrespective of thepresence of cattle drinking points. These dataindicate that cattle drinking access points wereunlikely to further impact negatively on waterquality in streams of water quality of Q3 or less.Complete exclusion of cattle from watercourses

might not be cost- or environmentally effective insome intensive agricultural grassland systemswhere water quality is already at low levels (≤ Q3).

AcknowledgmentsThis project was funded by the Research StimulusFund of the Department of Agriculture, Fisheriesand Food under the National Development Plan2006.

ReferencesChadd, R. & Extence, C. (2004) The conservationof freshwater macroinvertebrate populations: acommunity based classification scheme. AquaticConservation: Marine and Freshwater Ecosystems14: 597-624.

Council of the European Union. (2000) WaterFramework Directive 2000/60/EC establishing aframework for community action in the field ofwater policy. Official Journal of the EuropeanCommunities L 327/1.

Department of Agriculture, Food and Forestry.(2007) The Rural Environment Protection Scheme.Government Publications Office, Dublin.

Maitland, P.S. (1979) Synoptic Limnology: theanalysis of British freshwater ecosystems. Instituteof Terrestrial Ecology.

Hay meadow plant communities on theShannon Callows: responses to summerflooding and changes in management

C.A. Maher1, 2, M.J. Sheehy Skeffington1 and M.J.Gormally2

1Botany and Plant Science, National University ofIreland Galway, 2 Applied Ecology Unit, NationalUniversity of Ireland Galway.Email: caitrionamaher@gmail.com

IntroductionFloodplain grasslands are composed of mosaics ofplant communities, the species diversity andcomposition of which are determined primarily bythe hydrological regime (Grevillot et al., 1998).Agricultural practices are the next most influentialfactor affecting plant communities (Grevillot et al.,1998). The Shannon Callow hay meadows havebeen managed in the same way for centuries andfarmers have valued the annual hay cut as anutrient-rich food source for their livestock. Theannual flooding has restricted agriculturalintensification but the waters and sediments bringnutrients to the meadows each winter. Eachsummer the meadows are cut for hay and thisannual removal of biomass plays an essential rolein maintaining the nutrient balance on themeadows and promoting floristic diversity. Recentsummer floods have brought additional nutrientsand prevented farmers from cutting the meadows.Here, we investigate the effects that a combinationof both summer flooding and lack of cutting havehad on the plant communities.

Materials and MethodsStudy areaThe Shannon Callow SAC covers 5,788 ha of wetmeadows and pastures. The designated areaincludes two Priority Annex 1 habitats: Moliniameadows and lowland hay meadows. The callowscomprise the most extensive river meadows inIreland and support bird species and populations ofboth national and international importance.

MethodologyEight hay meadows were selected between Athloneand Banagher. Forty-one relevés were recorded in2007 across 15 sample areas (each chosen tocomprise a homogenous plant community and toreflect the range of zones in relation to floodduration). Each relevé was recorded within a 2m x2m quadrat and repeated in 2010. The elevation ofeach quadrat was recorded using a differentialglobal positioning system. The elevation readingscombined with twenty years of river level data(1990-2009) was used to calculate a variety offlood variables. Landowners completedquestionnaires to ascertain details regardingmanagement on the meadows.

Results and DiscussionTable 1 summarises the flood and managementvariables for each of the sample areas (listed 1-15in order of increasing hydroperiod). The annualhydroperiod here refers to the total number of daysa sample area was submerged in a year and themean was calculated for 1990-2009. The summerhydroperiod indicates the number of days a samplearea was submerged from May to September(inclusive), calculated for 1990-2009. The meanhydroperiod for recent summers (2006-2009, incl.)was also calculated to highlight the recentincreased duration of hydroperiods.

Table 1. Summary of variables for each of the 15sample sites. The mean refers to the arithmeticmean in each case. ’90-’09 = 1990-2009; ’06-’09 =2006-2009.

Site’90-‘09 mean

annualhydroperiod

(days)

’90-‘09 meansummer

hydroperiod(days)

’06-‘09 meansummer

hydroperiod(days)

No. ofyears

not cut

1 21.0 0.0 0.0 12 40.7 0.0 0.0 23 51.3 0.3 0.8 34 72.9 2.5 7.3 25 74.8 2.4 7.1 36 85.2 3.7 10.5 27 103.4 5.8 16.5 38 119.0 4.3 21.7 09 137.2 10.7 26.3 110 146.0 15.5 39.0 111 148.6 13.6 33.8 312 151.9 14.3 35.7 213 159.7 16.9 41.9 014 163.6 17.1 42.9 315 168.7 18.5 43.8 1

Over half the sample areas were flooded for morethan 100 days per annum. Recent summer floodslasted over twice as long as the long term average.The last column on Table 1 shows the number ofyears each sample area was left uncut during 2006-2009, i.e. number of recent summer cuts missed.These meadows were not cut either because themeadow was flooded or, as was the case forsample areas 1 & 2, the area was inaccessible dueto the surrounding floods. For other sample areas(e.g. 8 & 13), summer flooding occurred, but thefarmer did cut hay when meadows were unflooded.

For most sample areas, a significant decrease inplant species richness was observed in 2010 whencompared with 2007 (Fig. 1). The greatest drop inspecies richness was observed for sample areas 2 -6 inclusive. These areas have annual hydroperiodsof less than 100 days and were left uncut for atleast two years. Those meadows which normallyexperience longer flood duration comprise wetterplant communities which are adapted to tolerateflooding and thus reacted less dramatically to thesummer flooding.

0

5

10

15

20

25

30

35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

2007

2010

Sample Area

Me

anSp

eci

esR

ich

ne

ss

Fig. 1. Mean plant species richness in 2007 and in2010 for sample areas 1-15. Error bars indicate SD.

A significant drop in β-diversity, as measured by the Shannon Diversity Index (Fig. 2), was alsoobserved for the relevés sampled in 2010 whencompared to relevés from 2007. This drop in β-diversity occurred across all sample areas, whetherthey were only flooded in the summers, only leftuncut (& not flooded) or both flooded in thesummers and not cut.

Sample Area 151413121110987654321151413121110987654321

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Sh

an

no

nD

ive

rsit

yIn

de

x

2007

2010

Year

Shannon Diversity Index in 2007 and 2010

Fig. 2: Shannon Diversity Index for the 15 plantcommunities in 2007 and 2010.

The drop in β-diversity (Fig. 1) is more significant than the drop in species richness (Fig. 2). Thisreflects the fact that many species recorded in 2007although still present in 2010 were present in muchlower abundances. A contrasting increase in theabundance of faster-growing species was alsoobserved. In particular, meadowsweet Filipendulaulmaria, which grew to over 1.5 metres in heighton some meadows, out-competed smaller forbs andeven vigorous grasses.

A non-metric multidimensional scaling (NMS)ordination of the relevés recorded in 2007 and2010 (Fig. 3) shows that relevés taken in 2010clustered to the top right of the graph and relevésrecorded in 2007 to the bottom left. Meanhydroperiod is strongly positively correlated withAxis 1 where as species richness and the diversityindices are strongly negatively correlated with thisaxis. Number of years since the meadow was cut isstrongly positively correlated with Axis 2. Thus,whereas the 2007 relevés are correlated with

species richness and diversity the 2010 relevés arecorrelated with longer flood duration and lack ofcutting.

Species Richness

Shannon Diversity

Total % cover

Simpsons Diversity

Annual Hydroperiod

Summer Hydroperiod

No. years uncut

Axis 1

Axis

2

Fig. 3. NMS Ordination of relevés taken in 2007open triangles) and 2010 (closed triangles). Axis 1explains 37.6% and Axis 2 explains 33.5% of thevariance in the data.

ConclusionsIt appears that a combination of consecutivesummer floods and lack of hay cutting hassignificantly reduced the plant species richness anddiversity on the callows. The summer floods havefacilitated the growth of a few dominant specieswhile retarding the growth of other species. Theplant communities which were not flooded duringthe summers of 2006-2009 also showed asignificant drop in diversity corresponding to alack of cutting. This highlights the need for anannual hay-cut on the meadows both as a means ofremoving nutrients and controlling fast-growingspecies. The annual cutting on hay meadows isessential not only for the meadow plantcommunities but also to maintain the vegetationstructure favoured by ground nesting birds, mostnotably, the corncrake. Although delaying haycutting until after the breeding season for groundnesting birds is preferable, to ensure habitat qualityit is not advisable to delay cutting beyond July orAugust (with timing depending on the hydrologyof the meadow).

AcknowledgmentsMany thanks to BirdWatch Ireland for all theirhelp, to the OPW and ESB for providing river leveldata. A special thanks to the farmers for allowingaccess and giving information on farming practice.

ReferencesGrevillot, F., Krebs, L. & Muller, S. (1998)Comparative importance and interference ofhydrological conditions and soil nutrient gradientsin floristic biodiversity in flood meadows.Biodiversity and Conservation 7: 1495-1520.

Development and design of locally targetedHNV programmes in Ireland-The Aran IslandsCase Study

P. McGurn1 and J. Moran2

1European Forum on Nature Conservation andPastoralism. 2 Institute of Technology, SligoEmail: patrick@efncp.org

IntroductionThe three islands of Aran, Inis Mór, Inis Meáinand Inis Oírr are a geological extension of thekarstic limestone region of the Burren in north Co.Clare. Agriculture is and has been an importantpart of the islands economy and the agriculturalpractises have created a High Nature Value(HNV) system. The area contains a mixture ofrare Irish and European habitat types includingorchid-rich grassland/calcareous grassland,lowland hay meadows, limestone pavement (Plate1) and machair. Over 75% of the total land area isdesignated as a Special Area of Conservation(SAC) under the EU Habitats Directive. Some ofthese priority habitats are now in bad condition(NPWS 2008) (Plate1).

Plate 1. Calcareous grasslands on Inis Mórshowing scrub encroachment due to changes intraditional agricultural practices.

The BurrenLIFE project has shown that byworking closely with farmers it is possible toimprove the condition of priority habitats throughlocally targeted solutions. This model could beextended to other HNV areas such as the AranIslands.

BackgroundThe principal farming enterprises on the AranIslands are single suckling production and sheep.Most of the cattle are sold as stores to cattledealers on the mainland. In 2000, the area farmedwas recorded as 3,025ha across 224 farmers onthe three islands, indicating an average size ofapproximately 13.5ha, significantly below thenational average of 31.4ha (CSO, 2000). Kelly(2008) analysed the structure of farms under

agreement in the Rural Environmental ProtectionScheme (REPS). Whilst the average farm size inREPS was 17ha, 29% were less than 10ha andonly 5% of the farms exceeded 40ha. 79% of thefarms were less than 20ha. Stocking rates werelow with 60% of the farmers having a stockingrate of less than 0.6 livestock units/ha. Manyfarms are very fragmented with some spread over12 separate plots of land (Smith et al., 2010). Thesmall fragmented farms coupled with lowstocking rates means the farms are on a pooreconomic footing. Resulting changes in theislands traditional agricultural system is affectingthe condition of many habitats leading to apotential overall loss in biodiversity.

To date the main policy instrument for targetingbiodiversity on individual farms within Europehas been agri-environment schemes, which inIreland has been implemented through REPS andmore recently the Agri-Environment OptionScheme (AEOS). Kelly (2008) carried out a studyof REPS on the Aran Islands. He concluded thatoverall, REPS had been a beneficial scheme to theislands, with 88% of the farmers participating.The study found that REPS had improvedknowledge of the environment within the farmingcommunity and had helped in the maintenance ofstonewalls that are so characteristic of the islandslandscape. However, the study exposed someimportant limitations in the context of the AranIslands. There was a lack of positive managementwith some specific conservation issues that theREPS scheme did not address. He concluded thata higher-tiered agri-environment scheme ormeasure is required to focus on the specifichabitat, species and cultural conservation issues,which should complement rather than replace thework being done by the REPS. REPS hasbenefited the islands through the provision ofeconomic support to farmers, but it has failed toadequately address a number of conservationissues, such as declines in cereal and vegetabletillage and associated rare arable plants and scrubencroachment on priority habitats designatedunder the EU Habitats Directive (Smith et al.,2010). Some of these shortcomings arose from thefact that REPS was a national-scale scheme andalthough adaptable to some regional-specificissues, Smith at al. (2010) found that it was notideally suited to meeting all ecological andagricultural needs in the Aran Islands.

The Burren approachFarmers and researchers managing similarlimestone grasslands in the Burren, Co Clare alsofound that the generic farm managementprescriptions under REPS were not particularlyeffective at conserving unique features of the

Burren landscape. This shortfall led to theBurrenLIFE initiative, an EU funded LIFEprogram that undertook work based on “on farm”research with 20 local farmers. It combined localfarming knowledge with best scientific practicesto develop a regionally targeted farming forconservation programme. A variety ofmanagement options for grazing, feeding andscrub control were researched. This combinedwith monitoring, education and awareness raisingprogrammes resulted in successful measures beingexpanded to the wider farming community. Theresult has been a locally targeted agri-environment scheme for a High Nature Value(HNV) area now known as the Burren Farmingfor Conservation Programme (BFCP).

Extending the BFCP to the Aran Islands.The similarities between the Burren and the Aranislands means much of research outcomes couldbe transferred to the Aran Islands, with somechanges to reflect the differences in the two areas.A number of Aran farmers took part in a studytour to the BurrenLIFE project in 2008. The AranIslands were also included as a case study area ina recent Heritage Council Study on High NatureValue farmland in Ireland (Smith et al., 2010),part of which included a series of publicconsultation meetings. The Heritage Council thenteamed up with the European Forum on NatureConservation and Pastoralism (EFNCP) to employa HNV officer for Ireland who would further theHNV work on the Aran islands and other areas.The work began with a series of workshops heldon each of the three islands in August 2010. Theseworkshops explored the HNV concept and aimedto initiate community participation. The BFCPwas discussed, identifying areas where it may bedirectly transferable or where possibleamendments were required. A total of 48 islandersattended the workshops with 17 peoplevolunteering as contact points to assist with thefuture developments of the project. One of themain differences highlighted between the AranIslands and the Burren was farm structure withsmaller farm sizes and fragmented holdings on theislands. Overall the local community saw merit inexploring the BurrenLIFE model in further detail.A further series of meetings were held in lateSeptember 2010 in conjunction with REPStraining courses to present the combined resultsfrom the August meetings and collect anyadditional information or other views. Themeetings also offered an opportunity to highlightthe importance of the islands for natureconservation and the importance of agriculture inmaintaining these areas. Some of the farmersrequested an additional study tour to the BFCP,which took place in November 2010.

From the workshops, feedback from the additionalmeetings and farm visits, the project team inconsultation with local farmers drew up a list ofissues, proposed solutions and initial costings.The framework for an Aran Farming forConservation Programme based on the BFCPmodel was produced. Through the farmerrepresentative group, a further series of meetingswas held on each island and the programme wasexplained with knowledge gaps identified. On-going work involves the gathering of managementdetails from the farmers so best practices can bedocumented and the identification of trial sites fornecessary management works. The next step inthe process is to identify suitable funding sourcesto implement the plans either through LIFEprograms or through future changes in theCommon Agricultural Policy (CAP) reform.

DiscussionWhilst the Aran Islands and the Burren are similarlandscapes, the approach can be followed in arange of HNV landscapes to produce targetedprogrammes that meet the needs of thecommunity and improve the status of manypriority habitats. To be successful they require notonly sufficient funding but relevant research,dedicated individuals within the farmingcommunity and an understanding of the importantenvironmental services that their farming areaoffers as well as the production of quality food.Enhancement of environmental quality andsustaining the ecosystem services provided byHNV areas can only be achieved through theactive involvement of local communities. Thesecommunities are committed to the development oftheir areas. Through support from andengagement with other stakeholder much progresscan be made in developing and implementingtargeted programmes for the maintenance andenhancement of HNV landscapes.

ReferencesCentral Statistics Office (CSO) (2000) Census ofAgriculture. Stationery Office. Dublin.Kelly, I. (2008) A study of the Rural EnvironmentProtection Scheme on the Aran Islands. M.Sc.Thesis, University College Dublin.NPWS (2008) The Status of EU ProtectedHabitats and Species in Ireland. Department ofthe Environment, Heritage and LocalGovernment.Smith, G.F., Bligh, J., Delaney, E., Egan, M.,O’Donavan, G., O’Donaghue, P., O’Hara, K.(2010) Case Studies on High Nature ValueFarming in Ireland: Aran Islands andConnemara. A report to The Heritage Council,Kilkenny, Ireland.

Conserving Twite Carduelis flavirostris inIreland

D.T. McLoughlin1 and D.W. Norriss2

1Department of Environmental Science, Institute ofTechnology, Sligo. 2National Parks and WildlifeService, 7 Ely Place, Dublin 2.Email: mcloughlin.derek@itsligo.ie

IntroductionThe Twite Carduelis flavirostris is a member ofthe finch family that breeds and winters in Ireland.Although never common here, 100 years ago itwas believed to breed in all coastal counties. In thepast thirty years a considerable decline inpopulations has been observed and the latestestimate is less than 100 breeding pairs(McLoughlin & Cotton 2008). This populationbreeds in only five of the 32 counties on the islandof Ireland with over 85% of these occurring innorth Co. Mayo and west Co. Donegal. Thenational winter population is estimated at between650 – 1100 birds. Consequently, the Twite is oneof only three passerine species on the Red List ofBirds of Conservation Concern in Ireland (Lynas etal., 2007). Breeding Twite can be categorised asbeing ‘Endangered’ using the International Unionfor the Conservation of Nature (IUCN) criteria forthe categorisation of Red List species and are thusconsidered to be facing a ‘very high risk ofextinction in the wild’ in Ireland.

This paper presents an overview of the mainelements of the Twite’s ecology in Ireland with asummary of the primary actions required toconserve Twite as a breeding species in Ireland.

MethodsMovement patternsData on the movement patterns of Twite in the twobreeding strongholds was generated by monitoringcolour-ringed birds in the breeding and winterseasons. Radio tracking was also used to analysehome-range size and dynamics. This is presentedin McLoughlin et al. (2010).

Habitat requirementsHabitat selection studies focused on the breedingpopulations in north Mayo and west Donegal.Walkover surveys and radio tracking were used todetermine habitat preferences of Twite within 2 kmof their breeding colonies. Habitats were classifiedusing the Fossitt (2000) classification. Resultswere analysed using compositional analysis(McLoughlin 2009). The assessment of habitatrequirements during the winter season was basedlargely on evidence from similar studies inScotland and England e.g. RSPB Twite RecoveryProject (Gowthorpe 2009).

Results and DiscussionMovement patternsIn the course of this study, 492 birds were ringedof which 480 were caught outside the breedingseason; 57 (12%) were resighted on their breedinggrounds. The breeding birds spent most of thewinter season within 28 km of their breeding areas.Two birds ringed in the study areas during thewinter season were subsequently resighted duringthe breeding season on Islay and Mull of Kintyrerespectively. These results suggest that Irishbreeding Twite are mainly sedentary and thatpopulations appear to be augmented by Scottishbreeding birds during the winter months.

Breeding season foraging and nesting areasForaging Twite selected lower saltmarsh habitatsin west Donegal and dry-humid acid grasslandhabitats in north Mayo, where lower saltmarsh didnot occur. Wildflowers along tracks and roadswere important in both areas in April and earlyMay. The sward height in most of the foragingareas tended to be <100 mm, which, depending onthe target plant, allows access for birds to feed onthe seed of shorter plants, e.g. Pearlwort Saginanodosa. This shows the importance of grazing inthe foraging habitat during the breeding season tomaintain suitable sward height and to prevent adominance of a rank sward.

Of 72 nest sites found between 2005 and 2011 inDonegal and Mayo, 68 occurred in long HeatherCalluna vulgaris with only four using patches ofBracken Pteridium aquilinum. Twite stronglydepend on long heather with a heterogeneous mixof moorland vegetation for roosting and nesting.This highlights the importance of maintaining andincreasing the extent of long heather adjacent tobreeding colonies. It should be noted that at theMayo study area this vegetation type constitutedapproximately 1% of their home range. Thesenesting areas should ideally be <2.5 km from lowintensity agricultural lands where the Twite forage.

Targeted food plantsThe target food plants of Twite throughout thebreeding season comprise solely of common plantsthat produce relatively small seeds includingDandelion Taraxacum agg., Chickweed Stellariamedia, Sea Plantain Plantago maritima, ThistleCircium spp., and Autumn Hawkbit Leontodonautumnalis.

Winter seasonAlthough the winter ecology of Twite has not beenstudied in Ireland, it is thought that there may be abottleneck in the availability of suitable seedtowards the end of the winter season (Raine 2006).

Based on successful projects in Scotland,McLoughlin (2011) suggests several options thatwould provide suitable seed, including RadishRaphanus sativus and Turnip Brassica rapa,throughout the winter season. Seed prescriptionsfor current wild bird plots (e.g. LINNET plots inREPS) tend to be too large for Twite.

Conservation implications and recommendationsThe implementation of successful conservationplans for bird species can often be complicated bythe large areas the species may cover betweenwinter and summer seasons. In the case of twitehowever, due to the sedentary nature of ourpopulation, conservation action plans focused inthe areas they occur have the potential to be highlysuccessful for breeding and wintering populations.

A document giving guidance on the enhancementof Twite habitat through suitable land managementhas been prepared (McLoughlin 2011) and iscurrently available to land managers and theiradvisors in areas where Twite occur.

Table 1. Summary of suggested actions forfarmland conservation of Twite

Primary suggested actions (April-October)

1. Maintain, or create, a heterogeneous mix ofmoorland vegetation, particularly long Heather.2. Maintain bracken stands but prevent them fromincreasing in dominance on dwarf shrubs.3. Prevent afforestation of potential breeding areasby careful consideration of the location ofproposed plantations.4. Avoid topping of Thistles and Sorrel Rumexacetosa in potential feeding areas.5. Avoid the chemical spraying of wildflowers.6. Avoid agricultural 'improvement' of enclosedfields within a 3km radius of breeding areas.7. Maintain meadows with seeding wildflowersthroughout the breeding season.8. Restoration of improved, flower-poor fields totraditionally managed hay meadows.

ConclusionsIn the years prior to this project, a serious lack ofknowledge of the ecology of Twite in Ireland wasa major barrier to designing and implementingconservation measures to secure their future as abreeding species. Today however, we have adetailed knowledge of many aspects of theirecological requirements and we have acomprehensive plan (McLoughlin 2011) to addressthe issue of their possible extinction here.

Due to the precarious situation regarding the Twitebreeding population in Ireland, it is vital that landmanagement prescriptions and species policy nowfocus on their conservation. As Twite are not

Annexed species under the EU Birds Directive,they have not been included in any previous agri-environment schemes.

It is crucial that we incorporate species such asTwite in future agri-environment policy bytargeting specific measures on a bio-geographicalbasis (e.g. High Nature Value farmland policy).Implementation of these measures would alsobenefit a host of other flora and fauna in theseareas e.g. Red Grouse Lagopus lagopus. Withoutsuch measures it is most likely that Twite willfollow the fate of Corn Bunting Emberizacalandra, another farmland bird which becameextinct in Ireland in the early 1990s.

AcknowledgmentsThis research has primarily been funded by NPWSbut also benefited from funding under TheHeritage Council Wildlife Grants Scheme 2010.Sincere thanks to all who contributed data andadvice to this project. We are also indebted to themembers of the farming community who kindlypermitted us access to their lands.

ReferencesFossitt, J.A. (2000) A Guide to Habitats in Ireland.The Heritage Council, Kilkenny.

Gowthorpe, K. (2009) Farmland Advisers’ twiteHabitat Intervention Handbook. RSPB, UK.

Lynas, P., Newton, S.F. & Robinson, J.A. (2007)The status of birds in Ireland: an analysis ofconservation concern 2008-2013, Irish Birds 8:149-166.

McLoughlin, D. & Cotton, D. (2008) The Status ofTwite Carduelis flavirostris in Ireland 2008. IrishBirds 8: 323-330.

McLoughlin, D. (2009) The Ecology of TwiteCarduelis flavirostris in Ireland. Unpublished PhD,Institute of Technology, Sligo

McLoughlin, D., Benson, C., Williams, B., &Cotton, D. (2010) The movement patterns of twopopulations of Twites Carduelis flavirostris inIreland. Ringing and Migration, 25: 15 –21.

McLoughlin, D.T. (2011) Managementprescriptions for Twite in Ireland. Irish WildlifeManuals, No. 52. NPWS, Dept. of Environment,Heritage and Local Government, Dublin, Ireland.

Raine, A.F. (2006) The breeding ecology of TwiteCarduelis flavirostris and the effects of uplandagricultural intensification. PhD thesis, Universityof East Anglia.

How does animal stocking rate influenceagricultural grassland biodiversity whenmeasured at different scales?

B.J. McMahon1, A. Anderson1, T. Carnus1, A.J.Helden2, H. Sheridan1 and G. Purvis1

1UCD School of Agriculture, Food Science andVeterinary Medicine, University College Dublin,Belfield, Dublin 4, Ireland, 2Animal andEnvironmental Research Group, Department ofLife Sciences, Anglia Ruskin University, EastRoad, Cambridge CB1 1PT, UKEmail: barry.mcmahon@ucd.ie

IntroductionIndividual biological taxa operate at differentscales within agricultural ecosystems. As a resultthere is currently no consensus as to what is theoptimal scale to implement effective agri-environment policy (Gabriel et al., 2010). In orderfor indicators to be used to their fullest advantage,it is necessary to understand the ecologicalrelationships between the chosen indicatorgroup(s) and wider community structure, as well asthe particular ecological influences they reflect(Paoletti 1999). In this study, plants, parasitoids,birds and habitats were utilised as responseindicators to animal stocking rate as each of theseindicators operate at different scales withinagricultural ecosystems. The nature of the responseof these indicators provides important informationas to the targeting of different biological taxawithin agri-environment schemes.

Materials and MethodsPlants, parasitoid and farmscale bird surveys werecarried out on 119 grass-based farms located inthree regions in the Republic of Ireland during twofield seasons (2007-08 and 2008-09).

Plants were sampled using quadrats at the fieldmargin and in the field. Parasitoids were collectedfrom field swards were sampled using a VortisInsect Suction Sampler (Burkard ManufacturingCo Ltd, Rickmansworth, Hertfordshire, UK)(Arnold 1994). Standard yellow pan traps with awindow interceptor were used to collect mobileflying parasitoid populations. Bird surveys wereconducted using standardise techniques (Bibby etal., 2000; McMahon et al., 2008). In addition,habitats on each farm and within the surroundinglandscape were also classified and quantified.Animal stocking rates, calculated as livestock unitsper hectare (LU/ha) were used as an indicatormeasure of agricultural intensity.

AnalysesAnimal stocking rate, calculated as standardisedlivestock units per ha (LU/ha), was calculated as a

measure of overall agricultural intensity on thesurveyed farms, following the methodology of theIrish National Farm Survey (Anon. 2009).Generalized Linear Mixed Models were used toquantify the relationship between animal stockingrate and each of the indicators.

Results and DiscussionThe field scale indicators of plant species richness,parasitoid taxon richness and abundance werenegatively associated with animal stocking rate(Table 1). There was a positive relationshipbetween stocking rate and total winter bird speciesrichness and abundance, and the species richnessand abundance of winter Farmland Bird Indicatorspecies (Gregory et al. 2004). However, the overallrelationship was quadratic, indicating that above anupper limit intensity has a negative influence(Table 1). All other indicators showed no responseto animal stocking rate.

Table 1. Summary of likelihood ratio tests (χ2)with the effect of stocking rate on the selectedindicators.

IndicatorStocking Rate(χ2, P value)

Effect ofstocking

rateField plant speciesrichness

5.55, 0.019 Negative

Field parasitoidtaxon richness

5.15, 0.023 Negative

Field parasitoidabundance

3.36, 0.067 Negative

Winter birdsspecies richness

4.56, 0.033 Quadratic

Winter birdabundance

15.85, <0.001 Quadratic

Winter FarmlandIndicators speciesrichness

8.55, 0.003 Quadratic

Winter FarmlandIndicatorsabundance

16.23, <0.001 Quadratic

ConclusionsThe lack of a consistent response from theindicators to variations in stocking rate indicatesthat maximising biodiversity within agriculturalecosystems requires taxa to be targeted at theirappropriate operational scale. A variation in taxaresponse to farm system has also been reported in aprevious study (McMahon et al., 2010). It is notsurprising that our data revealed a significantlynegative influence of stocking rate on swardspecies richness in the centre of surveyed fieldsand the abundance and diversity of parasitoidwithin the sward; the latter group beingparticularly good indicators of taxon richness ofwider arthropod populations within agricultural

grasslands (Anderson et al., 2011). Theexplanation for this is that increased nutrient inputlevels can have a marked influence on both swardplant and arthropod communities in grasslands,with a generally negative effect on speciesrichness.

The positive effect of stocking rate (and byinference a positive influence of overallmanagement intensity within managed grasslandfields) on winter bird populations is counter-intuitive and contradicts any assumption thatgrassland management intensity has a necessarilynegative impact on all aspects of farmlandbiodiversity. A partial explanation for winter birdpopulations, including Farmland Indicator species,occurring in greater numbers on intensivelymanaged fields is that soil invertebrates, especiallyearthworms can be significantly more abundant (ifnot more diverse) under conditions of greaternutrient input levels (Curry et al., 2008).

The results of this study provide two importantinsights that should inform agri-environmentpolicy in the future, within Ireland but alsothroughout the EU.

Firstly, for development of optimally customisedagri-environment schemes where the aim is theprotection or enhancement of biodiversity, thescale of implementation is vital and should dependon the taxa that are targeted. The results of thisstudy and Gabriel et al. (2010) provide substantialevidence to support this perspective.

Secondly, the presence of relatively intensivegrassland management has a place in agricultureand may present an opportunity so long as it is notcoupled with widespread removal of non-croppedhabitat.

Recognition of these is essential to facilitate theeffective implementation of agri-environmentmeasures to maximise biodiversity within Irish andEU farmland.

AcknowledgmentsFunding was provided by the Irish Dept. ofAgriculture Fisheries and Food Research StimulusFund 2006 (06 0376). The co-operation of farmerswho provided sites was greatly appreciated.

ReferencesAnderson, A., McCormack, S., Helden, A.,Sheridan, H., Kinsella, A. & Purvis, G. (2011) Thepotential of parasitoid Hymenoptera asbioindicators of arthropod diversity in agriculturalgrasslands. J. App. Ecol. 48, 382–390.

Anon. (2009) National Farm Survey 2007 Data.Teagasc (Irish Agriculture & Food DevelopmentAuthority), (available at URL:issda.ucd.ie/documentation/Usermanual.doc,accessed 16 March 2010)

Arnold, A.J. (1994) Insect suction samplingwithout nets, bags or filters. Crop Prot. 13: 73-76.

Curry, J.P., Doherty, P., Purvis, G. and Schmidt,O. (2008) Relations between earthwormpopulations and management intensity in cattle-grazed pastures in Ireland. Appl Soil Ecol. 39: 58-64.

Gabriel, D., Sait, S.M., Hodgson, J.A., Schmutz,U., Kunin, W.E., and Benton, T.G. (2010) Scalematters: the impact on biodiversity at differentspatial scales. Ecol. Lett. 13: 858–869.

Gregory, R. D., Noble, D. G. and Custance, J.(2004) The state of play of farmland birds:population trends and conservation status oflowland farmland birds in the United Kingdom.Ibis 146 (Suppl. 2): 1–13.

McMahon, B.J., Purvis, G. and Whelan, J. (2008)The influence of habitat heterogeneity on birddiversity in Irish farmland. Biol. Environ 108B: 1-8.

McMahon, B.J., Helden, A., Anderson, A.,Sheridan, H., Kinsella, A. and Purvis, G. (2010)Interactions between livestock systems andbiodiversity in South-East Ireland. Agri. Ecosys.Environ. 139:232-238.

Paoletti, M.G. (1999) Using bioindicators based onbiodiversity to assess landscape sustainability.Agri. Ecosys. Environ. 74: 1–18.

Lethal and sub-lethal effects of ivermectin ontwo common species of dung beetle in alaboratory experiment

N.M. O’Hea1, 2, L. Kirwan1, 3, P.S. Giller2 and J.A.Finn1

1Teagasc, Environment Research Centre,Johnstown Castle, Wexford, Ireland, 2Departmentof Zoology, Ecology and Plant Science, UniversityCollege Cork, Ireland, 3Waterford Institute ofTechnology, Co. Waterford, IrelandEmail: john.finn@teagasc.ie

IntroductionDung decomposition is an important ecosystemservice in grazed grasslands and underpinsefficient nutrient cycling. Dung beetles play animportant role in dung decomposition. In addition,dung beetles constitute part of the diet of severalvertebrate wildlife species, including bats and birdsof priority conservation interest.Ivermectin is an anthelmintic veterinary medicine,and is excreted in the dung of treated livestock in amainly unmetabolised form. Ivermectin is knownto have toxic effects on dung beetles, but moststudies to date have been conducted on tropical andsub-tropical species. Relatively few laboratorystudies have focused on the specific effects ofivermectin on survival and development of northtemperate dung beetles.Susceptibility of dung beetles to the lethal and sub-lethal effects of ivermectin (and other relatedcompounds) in dung is of particular concern,because of the potential for reduced dung beetlebiodiversity, impaired dung decomposition andreduced prey resources for wildlife. Currentwildlife management guidelines of conservationauthorities (e.g. Natural England, Joint NatureConservation Committee) recommend livestockhusbandry practices that at least limit the use ofanthelmintics such as ivermectin in order toeliminate potential ecotoxicological risks forwildlife. Nevertheless, further evidence is desirableto support such recommendations.

We investigated the lethal and sub-lethal effects ofivermectin on different life history stages of twowidely distributed and abundant north temperatebeetle species. In this study, a series of bioassayswere conducted using two species that areabundant and have a widespread distribution innorth temperate areas i.e. Aphodius ater (de Geer)and A. rufipes (L.). We investigated the effect ofivermectin concentration on:a) survival of adult beetles,b) oviposition by adults,c) larval development rates andd) survival of larvae.

Materials and MethodsCattle were divided into four groups: an untreatedcontrol group and three treatment groups in whichanimals received a subcutaneous dose ofivermectin (Qualimec™) by injection (0.2 mg/kgbody weight). Following subcutaneous injection,ivermectin concentrations in dung typically reach apeak at 3-5 days post-treatment, and thereafterdecline to low detection limits. Thus, to vary theivermectin concentration in dung, the treatmentgroups were dosed on different days (i.e. at 7, 5and 3 days prior to dung collection) but dung wascollected from all four groups (including a controlgroup) on the same day to eliminate anydifferences in dung quality (e.g. dung moisturecontent) which might arise if dung was collectedon different days. Dung was collected separatelyfrom all groups.

Four bioassays were carried out for each beetlespecies using two dung types (cattle and calf),giving eight bioassays in total (Table 1). Groups ofadult beetles were initially added to replicate dungpats from each experimental group. Adult survivaland oviposition were measured, and replicateswere repeatedly inspected to determine larvaldevelopment and survival of the eggs laid by theadults. (See O’Hea et al. (2010) for details).

Generalised linear mixed models (GLMMs) wereused to assess the effect of ivermectinconcentration on beetle survival and development.In each analysis (a-e), fixed effects ofconcentration, dung type (calf/cattle dung), beetlespecies (A. ater/A. rufipes) and their interactionswere fitted. A random effect was incorporated toaccount for variation among bioassays. Thenumber of surviving adults (a), number of eggslaid by A. rufipes females (b), and number ofindividuals surviving at the end of the bioassay (e)were all modelled using Poisson regression(GLMM with a Poisson distribution and log linkfunction). The effect of ivermectin concentration,dung type, beetle species (A. ater/A. rufipes) andtheir interactions on the probability of reaching aparticular life stage by a certain time (analysis c)were assessed using an ordinal model (GLMMwith a multinomial distribution and a cumulativelogit link function). The proportional survival oflarvae (d) was modelled using logistic regressionwith binomial distribution and logit link. Allanalyses were fitted using the GLIMMIXprocedure in SAS.

Results and DiscussionThere was a highly significant and negative overalleffect of ivermectin on larval development (e.g.Fig. 1). For example, the predicted probability of alarval individual developing beyond instar III was

significantly affected by ivermectin for bothspecies. The largest effects occurred in thebioassays with A. ater in cattle dung. Theseindicated an 80% probability of A. ater larvaehaving developed beyond larval instar III after 4weeks in the dung without ivermectin, whereas thisprobability dropped to about 15% in dung with 0.2mg of ivermectin per kg (wet weight of dung) (Fig.1). There can be strong pressures on larvae tocomplete their development before conditions inthe dung pat become unsuitable, and additionaldelays to larval development by ivermectin mayincrease larval mortality.

Fig. 1. Example of effects of ivermectin on larvaldevelopment times of Aphodius ater. Ivermectinlevels of 0, 0.1 and 0.2 mg kg-1 are shown as short-dashed, long-dashed and continuous lines,respectively.

Increased ivermectin concentration consistentlyhad a highly significant negative effect on theabundance of surviving individuals at the end ofthe bioassays. Highest mean numbers of survivingnewly emerged adults (A. ater) or prepupae (A.rufipes) were found in the control dung pats withno ivermectin. In the majority of cases, there werefew, if any, survivors at the end of the study in thedung pats with highest ivermectin levels (Fig. 2).

Fig. 2. Example of effects of ivermectinconcentration on proportional survival of larvae ofA. ater. Plotted values are the final number ofindividuals as a proportion of the initial number ofeggs.

In general, ivermectin concentration did not havesignificant negative effects on adult survival (overa period of 4-10 days). The number of eggs perfemale A. rufipes was significantly reduced byivermectin concentration in one of two bioassays,but the magnitude of the effect was not large.

ConclusionsLarval development rates were significantlyslowed by ivermectin. Ivermectin had significantnegative effects on the survival of larvae. Overall,ivermectin concentration caused large andsignificant reductions in the cohort size from anindividual dung pat that would potentiallycontribute to the next generation of beetles.

Extrapolating from controlled experiments at thescale of individual pats to field conditions,however, invokes several factors that affect thelevels of ivermectin in dung pats, and the actualimpact on dung beetle populations and otherfarmland wildlife (see O’Hea et al. (2010) forfurther discussion). Given the variety of factorsinvolved across several scales, it is not surprisingthat there is considerable uncertainty about theextent to which dung beetle populations aredepleted by ivermectin usage, and about the knock-on effects on populations of vertebrate wildlife thatprey on dung beetles.

ReferencesO’Hea, N., Kirwan, L., Giller, P.S. and Finn, J.A.(2010) Lethal and sub-lethal effects of ivermectinon north temperate dung beetles, Aphodius aterand Aphodius rufipes (Coleoptera: Scarabaeidae).Insect Conservation and Diversity 3: 24–33.

Experts’ assessments of biodiversity optionsand supplementary measures in REPS 4

D. Ó hUallacháin1, J.A. Finn1, M. Gormally2, andC. Carlin 2

1Teagasc, Environment Research Centre,Johnstown Castle, Wexford, 2Applied EcologyUnit, NUI GalwayEmail: daire.ohuallachain@teagasc.ie

IntroductionAgri-environment schemes (AESs) in the EU area major contributor towards CAP objectives toreverse biodiversity decline, achieve good waterquality by 2015 and achieve the Kyoto targets formitigating climate change. Member States areobliged to implement monitoring and evaluationof their respective agri-environment programmes.The evaluation process is intended to identify theextent to which policy objectives are beingfulfilled, and to identify any changes necessary tobridge the gap between policy aims and outcomes.Summary reports on agri-environment policyevaluations have concluded that there has beeninsufficient measurement of the preciseenvironmental outcomes from agri-environmentschemes. Participation in AESs per se does notguarantee the actual delivery of environmentalprotection or improvement, and only themonitoring of actual performance andenvironmental outcomes can demonstrate the truevalue and environmental impacts of agri-environment schemes (Kapos et al., 2009). Aconsequence of the lack of environmentalmonitoring of schemes is their impaired ability toidentify either successes or failures, and to learnhow to improve their environmental effectiveness.In the absence of national-scale quantitative data(Finn and Ó hUallacháin, in press), the aim of thisstudy was to elicit experts’ judgements on theexpected environmental performance of selectedelements of REPS.

Materials and MethodsWe consulted with a group of eight Irish agri-environmental experts to assess the wildlife valueof current supplementary measures and options inthe REPS 4 scheme. The selection of experts wasbased on several criteria: knowledge andexperience of biodiversity, agri-environmentpolicy, applied agro-ecological research andapplied interpretation of REPS policy in advisingfarmers (Finn et al., 2009).The assessment utilised experts’ judgements ofthe effectiveness of the REPS options andsupplementary measures that are relevant tobiodiversity. The assessment occurred in twostages.

First, experts scored each option using a scoringscale for each of five criteria, as follows: validity of the cause-and-effect relationship

between the intended objective and theprescribed management,

degree of institutional implementation, degree of farmer compliance, the extent to which the measure achieved an

appropriate match between the distribution ofenvironmental issues and participation and,

the extent to which participation wassufficient to achieve the environmentalobjective.

Second, the scores were collated and the group ofexperts discussed each option, elaborated on thejustification for their decisions and aimed toachieve consensus.

Results and DiscussionThe majority of biodiversity options received highscores for both the cause-and-effect andcompliance criteria. Therefore, for the majority ofmeasures and options, correct implementation ofthe management prescriptions is expected toachieve the environmental objective.Nevertheless, many measures were consideredunlikely to be as effective as expected. Severaloptions were expected to have little or noenvironmental effect, and some of these wereassociated with medium to very high participationlevels. The assessment identified specific reasonswhy certain options were not expected to bewholly effective. Many options are likely to havelow or no effectiveness (at the scheme scale)because of insufficient participation levels (Figure1).

Participation

0

102030

40506070

Perc

en

tag

eD

istr

ibu

tio

n

Fig. 1REPS othe scoA scorachieve

The exobjectishouldThe ob

0 1 2 3 4 5

Scores

. Distribution of experts’ assessments ofptions and supplementary measures across

re categories for the participation criterion.e = 0 indicates insufficient participation to

the environmental objective.

perts recommended that the aims andves of the scheme and individual optionsbe stated with greater clarity and precision.jectives should clearly identify the type of

biodiversity to be benefited/ targeted, and betterexplain how this will be achieved by themanagement prescriptions.

Other recommendations were relevant to designand implementation choices at the scheme-scale: move away from a ‘one-size-fits-all’ approach

in favour of spatial targeting. consider the additional effectiveness that may

be achieved from spatial targeting orincentivised participation of groups of farmersat the ladscape scale, as well as recommendedminimum participation levels to achievespecific environmental objectives.

reduce the choice of measures within the agri-environment scheme. A tiered approach wasrecommended, with the choice of optionsbeing strongly guided toward those best suitedto the farm conditions.

The experts consistently emphasised a number ofother comments. Biodiversity and habitat conservation

objectives should be afforded higher priority,especially given that most agri-environmentfunding in Ireland is allocated to biodiversity.REPS should improve the provision of advicefor the protection and management of existinghabitats. For relevant habitats, there should bemeasures that target the achievement offavourable conservation status.

Clearly prioritise and distinguish among theneed for conservation of existing habitats,enhancement of degraded habitats, andcreation of new habitats.

Ensure greater alignment betweenbiodiversity objectives in REPS and those oflocal and national biodiversity priorities, withthe overall aim of achieving the renewed EUtarget of halting biodiversity loss by 2020.Guidance on the latter is provided by theNational Biodiversity Plan, relevant policydocuments and other publications e.g. RedList Data books, National Strategy for PlantConservation etc.

The issues raised by the experts would be relevantto any evaluation of the design and structure ofexisting measures that may be included in futureagri-environment schemes. However, highestpriority should go towards:

strengthening links between the biodiversityobjectives of REPS and national conservationpriorities. This will be necessary to meet thenew EU target of halting biodiversity loss by2020, and;

monitoring and measuring the environmentaleffectiveness of REPS. Where necessary,

monitoring would also help learn howimprove any measures with deficiencies.Most importantly, the aim of monitoringwould be to demonstrate the environmentalbenefit of well-designed measures andoptions.

ConclusionsThe use of expert groups proved to be an efficientand effective method to:

(i) assess the likely environmentaleffectiveness of biodiversity options;

(ii) identify specific aspects of options thatare in need of improvement, and;

(iii) highlight modifications that shouldimprove environmental effectiveness.

The use of experts’ judgements can be particularlyuseful as a method for achieving rapid feedbackon the design of new schemes or measures.

AcknowledgmentsFunding was provided under the NationalDevelopment Plan, through the Research StimulusFund, administered by the Department ofAgriculture, Fisheries and Food.

ReferencesFinn, J.A., Bartolini, F., Kurz, I., Bourke, D. andViaggi, D. (2009) Ex post environmentalevaluation of agri-environment schemes usingexperts’ judgements and multicriteria analysis.Journal of Environmental Management andPlanning 52: 717-737.

Finn, J.A. and Ó hUallacháin, D. A review ofevidence on the environmental impact of Ireland’sRural Environment Protection Scheme (REPS).Biology and Environment (in press).

Kapos, V., Balmford, A., Aveling, R., Bubb, P.,Carey, P., Entwistle, A., Hopkins, J., Mulliken, T.,Safford, R., Stattersfield, A., Walpole, M. andManica, A. (2009) Outcomes, not implementation,predict conservation success. Oryx, 43: 336-342.

HNV farming and the management ofupland biodiversity on the IveraghPeninsula, South Kerry.

Eileen O’RourkeDepartment of Geography, UCC, CorkEmail: e.orourke@ucc.ie

IntroductionThis short paper deals with issues surroundingfarming systems and the management ofpriority upland habitats on the IveraghPeninsula. We combined a comprehensivefarm management survey of 80 Iveragh hillsheep farms with habitat mapping and adetailed ecological survey on a subset of 21core farms. The management survey coveredissues relating to farming systems – grazingmanagement, stocking density, animal breeds,along with environmental practices and farmeconomics. Using ecological criteria, the farmswere classified as undergrazed, overgrazed orsustainable grazed.

Active management of farming systems, inparticular grazing management, is critical toupland biodiversity. Both under- and over-grazing adversely affect plant diversity. Overthe years, European CAP policy has had amajor influence on the management of theuplands, as we move from the high stockingdensities and overgrazing associated with theEuropean Ewe Premium in the 1980s, to theimplementation by the Irish Government of theCommonage Framework Plan in 1998 (whichbrought about a compulsory destocking of allcommonages), through to the 2005 SingleFarm Payment (SFP), which effectivelydecoupled income from production and led toa further destocking of the uplands.

ResultsOverall, grazing management on the Iveraghwas found to be extensive, but there wasvariability depending on site specific andproduction objectives. Mean annual stockingrates on semi-natural upland areas werecalculated to be 0.29 LU/ha. However, thisfigure masks the fact that at 0.48 LU/ha,stocking rates during the grazing season can besignificantly higher. Stocking density as anindication of grazing pressure is problematicas it gives no indication of the density offeeding. Furthermore, it does not take intoaccount the altitude at which grazing occurs,

the spatial distribution of the grazers, foragingbehaviour, impact of shepherding, breed andage of animals, burning regime, supplementaryfeeding, time of year, along with previousmanagement history. On the Iveragh Peninsulawe remarked that the same stocking rate maycause one area to be ecologically damaged,and yet leave another almost undisturbed.Consequently, we recommend that grazingcontrols should be locally specific and ahabitat specific stocking rate would be of moreuse than applying a blanket stocking density.Livestock grazed the uplands for an average of221 days, a significant reduction to thetraditional year round grazing regime. We alsoremarked that the decline in store cattlegrazing in the uplands, or the switch tosupplementary feed suckler cows, is associatedwith the current spread of bracken, gorse andhard rush. In keeping with current marketdemands for heavier lambs, the farm surveyfound that 42% of the traditional rusticBlackface ewes are now cross-breed withlowland breeds, resulting in different foragingbehaviour. The harsh conditions and lowquality of hill grazing resulted in low breedingsuccess rates of 0.6-0.8 lambs per ewe.

The mean surface area of the 80 farmssurveyed was 138 ha, and the average farmmanager’s age was 49 years. The hill farmstypically consisted of 59% upland, 20%improved ‘greenland’ and 21% rough land.Seventy percent of the farms had a share in anupland commonage, which constituted 32% ofthe area farmed (on average). Eighty-sixpercent of surveyed farms participated in theRural Environment Protection Scheme (REPS)and 52% of holdings were designated asSpecial Areas of Conservation (SAC). Ofsurveyed respondents, 73% had reclaimed landsince their time as farm operator. The resultant‘greenland’ allowed farmers to keep morelivestock on low ground for longer periods andthereby reducing farmers’ dependency onupland rough grazing and rustic sheep breeds.This observation combined with the fact thatthe majority of the farm managers are todayinvolved in off-farm work has led to an overallsimplification of the traditional managementstructure. We identified a trend towardsmoving farming down slope, concentrating thefarming system around the reclaimed‘greenland’, and the less intensive use of theupland rough grazing and commons

(O’Rourke and Kramm, 2009; Kramm et al.,2010). Over half the farmers surveyed noted asignificant increase in scrub over the last fiveyears in the uplands. We remarked a trendtowards intensification and extensification(even abandonment) on the same holding. Theecological data from the core 21 farms studiedfound that the majority (52.4%) wereclassified as undergrazed, 19% weresustainably grazed and 28.6% wereovergrazed.

The analysis of the labour structure on the 80farms found one Annual Work Unit (AWU)per 60 Livestock Unity (LU), a valuecompatible with low intensity farmingsystems. The farm management survey founda high incidence of pluriactivity on farms.Only 19% of farm households were solelydependent on their farms for a living. In theremaining cases, either the farm manager andor spouses had an additional income. Thirty-five percent of the surveyed farmers whoworked off farm worked in construction, asector that has subsequently seen a majordown turn. However, it is important to notethat even during the Celtic Tiger, farming stillconstituted 61% of family farm income. Eventhough the Iveragh Peninsula is a designatedarea of outstanding natural beauty, only 17%of the surveyed Iveragh hill farmers wereinvolved in tourism, and for those that were,tourism contributed less than 25% to thefamily farm income. Only 2.8% of ourrespondents saw farm multifunctionalism as aviable future, they have en mass opted forpluri-activity in the form of off farm work.Forty-nine percent of our respondentsindicated that they had no definite successor.However, 78% were confident that they wouldremain farming in the future, with the mainmotivation being attachment to the familyfarm, place, heritage and a way of life. Theyare not solely rational economic actors.

Another important finding from this researchis, despite the fact that the 80 hill farmers areoperating under similar environmental, policy,and market conditions, there is a markeddiversity within their farming styles, againwith biodiversity implications. A statisticallyrigorous Farm Typology produced fourdistinct farm types, which we labelled:Environmental Stewards (29 farms), SupportOptimisers (12 farms), Traditionalists (25

farms), and Production Maximisers (6 farms).The farms managed by the environmentalstewards had the highest amount ofsustainablymanaged upland grazing area, withthe least incidence of scrub encroachment. Thetraditionalists, who were less inclined to seethe environment as a valuable public good, hadthe heaviest stocking levels and the highestnumber of holdings classified as overgrazed.The support optimisers and the yieldmaximisers, the two extreme farming styles,had the highest incidence of scrubencroachment on their mountain grazing areasand the highest dependence on ‘greenland’.Overgrazed farms received similar levels ofREPS payments to sustainably grazed farms,with undergrazed farms receiving significantlyless. We also found that overgrazed farmsreceived the highest SFP per ha. Despite thefact that 83% of our respondents participate inREPS, their management styles, goals andattitudes to nature conservation differconsiderably. It seems unlikely that themaintenance of the Iveragh’s biodiversity-richupland habitats can be guaranteed by currentundifferentiated support programmes alone.This suggests the need for a more targeted andcustomised approach to financial support. Weconclude that effective policies for theconservation and management of farmlandbiodiversity requires a cross-sectionalapproach that can take account not only ofenvironmental criteria, but also the policyenvironment and land managers’ decisionmaking process & socio-economic objectives.

AcknowledgementsThis research was funded by ScienceFoundation Ireland (2006-2009). The otherUCC researchers involved in the project were:Nadine Kramm, Roslyn Anderson, JohnO’Halloran, Mark Emmerson, Nick Chisholm.

ReferencesKramm, N., Anderson, R., O’Rourke, E.,Emmerson, M., O’Halloran, J., Chisholm,N.(2010) Farming the Iveragh Uplands: A Taleof humans and nature, UCC, Cork, p. 128.

O’Rourke, E and Kramm, N. (2009) Changesin the management of the Irish Uplands: Acase-study from the Iveragh Peninsula.European Countryside, 1: 53-69.

Mapping the broad habitats of the Burrenusing satellite imagery

S. Parr1, J.A. Finn2 and G. O’Donovan3

1Burren Farming for Conservation Programme,Old School House, Carron, Co. Clare , 2Teagasc,Environment Research Centre Johnstown Castle,Wexford, 3Broadview Ecological Consultants, 2Broadview, St. Mary’s, Chalford, Gloucestershire,UK2Email: john.finn@teagasc.ie

IntroductionNational and international commitmentsincreasingly require decision-makers in the ruralenvironment to conserve and protect high-nature-value habitats, which are known to commonlyoccur on farmland and other areas throughoutIreland. But where exactly do such habitats occur?Often, there is surprisingly little knowledge on theextent and distribution of high-nature-valuehabitats, and such knowledge is an essentialprerequisite for informed decision-making. High-quality information on the quality and distributionof habitats allows more local-scale (e.g. countylevel or DED-level) prioritisation of habitats thatare of very high quality, or are severely threatened.In turn, such information can facilitate moretargeted conservation management. Traditionally,habitat mapping has occurred on a field-by-fieldbasis, which is very labour-intensive. Here, we usethe Burren as a case study area for theimplementation of a methodology that usessatellite remote sensing to identify different habitatand vegetation types.

Materials and MethodsThree cloud-free images were chosen for the study;Landsat 5 from May 1999 and Landsat 7 ETM+from April 2003 and August 2000. The imageswere processed in Erdas Imagine 8.6 and anunsupervised classification was carried out using10 classes to provide a baseline habitat map fortesting. Initial findings indicated that these latespring/early summer images were most suitable formapping improved grassland, species-richgrassland, limestone pavement, water bodies,urban/dwellings and dense scrub.The unsupervised classification was used to informthe spatial location of ground-truthing, whichinvolved habitat data being collected atapproximately 850 individual points. Based on theground-truthing data, a supervised classificationproduced 15 classes that clearly corresponded tobroad habitat types (full details in Parr et al.,2006). This study also described the grassland andheath vegetation of conservation interest in theBurren, and correlated the vegetation data with

environmental and management factors (see Parr etal., 2009).

Results and DiscussionThe final habitat map was created using a varietyof techniques as described above, in an attempt toimprove the level of accuracy obtained solely by asupervised classification. The methods used toclassify the habitats meant that the end result was acomposite of multiple images rather than a singleone containing all the relevant habitats. Asummary of the area (in percentage cover terms)associated with different broad habitats of theBurren is shown in Table 1.The following habitats were mapped (values inbrackets represent the percentage of the study areaoccupied by the habitat): unimproved grassland(31%) comprising 17% ‘strong winterage’ and14% ‘weaker winterage’; improved grasslands(28%); limestone pavement (20%) comprising10% bare limestone pavement and 10% part-vegetated limestone pavement; scrub (14%);Calluna heath and open scrub (3.4%), water bodies(1.5%) (incl. lakes, turloughs and lacustrinevegetation); dunes (0.2%) and; tillage (0.2%)(Table 1).Scrub was one of the least satisfactory classesdefined, mainly due to its confusion with shadowand some other classes. Scrub is an importantthreat in the Burren, so additional imageprocessing was carried out to try and improve itsaccuracy. A new image was produced bycombining the April 2003 and August 2000images. This resulted in improved separation ofscrub from other habitats. Scrub was estimated tocover 15-20% of the Burren karst region (Table 1).The greatest challenge is to distinguish betweenlow, open scrub and the limestone heaths of highconservation value that occur at higher altitudes.

ConclusionsThis map was one of the first to show thedistribution of the broad habitats of the Burren andwill be an important tool in aiding future decisionsas to how the habitats of the Burren should bemanaged to the benefit of both the farmer and theenvironment. The map provides the first estimateof the area of the Burren affected by scrubencroachment – this being one of the mostsignificant threats to the EU priority habitats in theregion. Full details in Parr et al. (2006, 2009).On a particularly challenging area with a highdiversity and complexity of habitats, remotesensing appears to offer a very effective and cost-efficient alternative to broad-scale habitat mappingon a field-by-field basis.The outputs provided by such mapping approachescould inform the targeting of agri-environmentalobjectives, and increase the efficiency of detecting

areas of high conservation value for monitoring bymore conventional methods.This study indicates that satellite imagery is a veryuseful tool for long term monitoring of habitats inthe Burren.

AcknowledgmentsThis project was made possible by an award fromthe Research Stimulus Fund of the Department ofAgriculture and Food, funded by the IrishGovernment under the National Development Plan2000 – 2006. We thank the farmers of the Burrenand the Burren IFA for their support and co-operation.

ReferencesParr, S., O’Donovan, G. and Finn, J.A. (2006)Mapping the Broad Habitats of the Burren usingSatellite Imagery RMIS 5190c End of ProjectReportwww.teagasc.ie/research/reports/environment/5190c/eopr5190c.htm

Parr, S., O’Donovan, G. Ward, S. and Finn, J.A.(2009) Vegetation analysis of upland Burrengrasslands of conservation interest. Biology andEnvironment, 109B: 11-33.

Table 1. Broad habitats identified in the study area using satellite imagery from about 2000).Broad habitat Area

(% cover)Description

Unimproved grassland

Strong ‘winterage’

Weaker ‘winterage’

31%

(17%)

(14%)

Lower productivity, species-rich, mostly winter grazed.

More productive, calcareous to neutral on deeper/clay soilsincluding - wet grassland, short Calluna heath & someImproved Grassland on thin soils.Less productive, calcareous, rocky, thin soils includingSesleria-dominated grassland and Dryas-dominated heath

Improved grassland 28% More intensively farmed - higher productivity, relativelyspecies-poor, including ‘rushy’ land

Limestone pavement (LP)

Bare LP Vegetated LP

20%

(10%)(10%)

Ranges from bare ‘massive’ limestone pavement throughisolated vegetated patches/bands to ~ 75% vegetated withvery thin soils

Scrub 14% Predominantly closed hazel but also whitethorn,blackthorn and holly scrub, including ash-hazel woodland.

Calluna heath & open scrub 3.4% These two habitats tended to overlap, but as guidance:1) high, exposed plateaux & north-facing steep slopes tendto be Calluna heath with tall, mature and/or senescentCalluna2) sheltered areas adjacent to existing scrub tend to be lowor open scrub

Greggan’s Wood shale 1.7% Conifer plantation, blanket bog, ‘rushy’ grassland on shaleWater/lacustrine/turlough 1.5% Includes water body and tall lacustrine vegetationDunes 0.2%Tillage 0.2% Underestimated, but small

The National Survey of Upland Habitats

P.M. Perrin1, J.R. Roche1, O.H. Daly1 and C.Douglas2

1BEC Consultants Ltd.,26 Upper Fitzwilliam St.,Dublin 2, 2National Parks and Wildlife Service(NPWS), 7 Ely Place, Dublin 2Email: pperrin@botanicalenvironmental.com

IntroductionThe uplands form Ireland’s largest expanses ofsemi-natural habitats and lie above the upper limitsof enclosed farmland. Almost 19% of Ireland isconsidered to support upland habitats includingblanket bog, heath, grassland, flushes, lakes,springs, exposed rock and scree. The conservationvalue of these areas is unquestionable, withnumerous EU Habitats Directive Annex I habitatsand many rare and threatened floral and faunalspecies being recorded there. Furthermore, over40% of the terrestrial area designated as candidateSpecial Area of Conservation (cSAC) in Irelandlies above 150 m in altitude. The vast majority ofthe uplands are actively farmed as marginalgrazing land, mainly for sheep, with large areasbeing held as commonage.

Loss and degradation of extensive areas of uplandhabitats increased since the introduction of forestrygrants and ewe headage payments in the 1980s andencroachment or intensification of other humanactivities including wind energy developments.Negative impacts include changes in plant speciescomposition, habitat fragmentation, drainage, soilerosion and, in some areas, landslides. Uplandhabitats may also be especially vulnerable toclimate change, including extreme weather events.

Sensitive, evidence-based land managementpolicies are needed to ensure that Annex I uplandhabitats maintain or attain favourable conservationstatus and to prevent the decline of rare orthreatened species. Towards this end, the NationalSurvey of Upland Habitats (NSUH) wascommissioned by NPWS and commenced in 2010following a pilot survey in 2009. The principalaims are to:

survey a representative sample of the full rangeof upland habitats in Ireland,

map the location, extent and condition ofhabitats recorded and to produce baseline maps,

map the distribution of rare and threatenedupland flora,

conduct a baseline conservation assessment ofAnnex I upland habitats and establish a seriesof geo-referenced plots for periodic monitoring,

devise a classification system for uplandvegetation based on analysis of relevés,

identify impacts, threats and trends especially inrelation to Annex I habitats.

Materials and MethodsA manual detailing the methodology to be used inthe NSUH has been produced (Perrin et al., 2010).

Survey sitesDuring 2009 and 2010, six sites covering 389.7km2 were comprehensively surveyed (Table 1).

Table 1. Sites surveyed during 2009 and 2010.Siteno.

Site Area(km2)

1 Mweelrea/Sheeffry/ErriffComplex cSAC

209.8

2 Corraun Plateau cSAC 38.93 Comeragh Mountains cSAC 62.94 Carlingford Mountain cSAC 31.05 Nephin Mountain 14.16 Croaghaun/Slievemore cSAC 33.0

Habitat mappingMuch of the uplands consist of intricateenvironmental gradients that reflect changes intopography, hydrology, soils and geology and arefar too complex to map separately in theconventional fashion. Hence the NSUH hasadopted an approach of mapping units that reflectconsistent habitat mosaics. Prior to field work,aerial photograph interpretation is used to digitisesurvey sites into high resolution polygons thatrepresent areas of consistent pattern or topography.

Field surveyors use ruggedized PDAs with GISsoftware and a real-time GPS location to navigateaccurately. Within each polygon, the percentagecover of each Fossitt (2000) and Annex I habitat isrecorded. A provisional, subjective uplandsclassification is used to record in more detail thevarious vegetation communities encountered.

Assessment of conservation statusWhilst all habitats encountered during fieldworkare mapped, the assessment procedure is focussedon 12 Annex I habitats that form the primary focusof the project (Table 2). Three aspects of thesehabitats are assessed: i) changes in area, ii)structure and function, through the recording of aseries of monitoring stops across the site, each ofwhich includes a relevé, iii) future prospects,through an examination of the intensity and trendsin land use and impacts.

Compilation of rare species dataWhilst the NSUH is not a rare species survey,existing records of rare species are being collatedwith new NSUH records on a site-by-site basis.

Production of a relevé-based classificationRelevé data recorded by the NSUH are beingcombined with existing relevé datasets from theIrish uplands (primarily PhD theses) to produce anobjective vegetation classification. To facilitatethis, relevés are also being recorded in non-assessment habitats. Multivariate statisticaltechniques including hierarchical cluster analysisand indicator species analysis are being employed.

Results and DiscussionHabitats and relevésThe main habitats mapped within the six sitessurveyed to date were 4010 Wet heath, *7130Active blanket bog and 4030 Dry heath,comprising 23.6%, 15.0% and 10.1% of the overallsurvey area. Neither 6150 Siliceous alpine andboreal grassland nor the upland ledge aspect ofhabitat 6430 had previously been recorded inIreland. Similarly, 6170 Alpine and subalpinecalcareous grassland was recorded for the first timeduring a reconnaissance survey of Ben Bulben.The approach to these three habitats is underreview. A total of 377 relevés have been recorded,including 339 in Annex I habitats (Table 2). Thereis considerable variation in species richness (alphadiversity) between habitats, but it is important tonote that many upland habitats support specialistspecies. Hence, even where species richness is low,there may be considerable contribution to the betaor gamma diversity of the farmland landscape.

Conservation statusThe most serious impacts recorded to date areovergrazing by sheep and peat erosion, whichresulted in the overall assessment of habitats 4010and *7130 as unfavourable at all sites. Whilst therehas been significant destocking of commonages inthe last decade, it appears that areas that arealready severely damaged are likely to continue toerode without practical intervention. At one site,widespread inappropriate heather burning resultedin the unfavourable assessment of habitat 4030.Rocky slope and scree habitats have largely beenassessed as favourable.

Rare and notable speciesNew county records were made for the clubmossDiphasiastrum alpinum (Co. Waterford) (Roche &Perrin, in press) and the mosses Andreaeamegistospora, Racomitrium affine and Polytrichumalpinum (Co. Louth). New stations were recordedfor notable species such as cowberry Vacciniumvitis-idaea and the moss Amphidium lapponicum.

Relevé-based classificationA dataset of 3,742 relevés has been used toproduce a provisional objective classification of sixmain groups divided into 63 vegetation types.

ConclusionsThe uplands contain some of Ireland’s mostbiodiverse and specialised habitats but sensitive,evidence-based farming and conservationmanagement practices are needed if we are to fulfilour obligations under the Habitats Directive andconserve this resource. The NSUH is providing thescientific data to underpin these policy decisions.

Table 2. Species richness in Annex I habitats fromthe six surveyed sites. N = sample size. R =richness. † indicates non-primary focus habitats.

Hab.code

Habitat name N R

4010 Northern Atlantic wet heaths 77 20.84030 European dry heaths 46 20.24060 Alpine and Boreal heaths 40 19.56150 Siliceous alpine and boreal

grasslands†12 16.6

*6230 Species-rich Nardus grasslands 9 30.36430 Hydrophilous tall herb fringe

communities†3 34.0

7130 Blanket bog (*active only) 73 19.57140 Transition mires 7 17.17150 Depressions on peat substrates

of the Rhynchosporion16 16.8

7230 Alkaline fens 12 23.58110 Siliceous scree of the montane

to snow levels21 18.5

8120 Calcareous and calcshist screesof the montane to alpine levels

0 -

8210 Calcareous rocky slopes withchasmophytic vegetation

2 27.0

8220 Siliceous rocky slopes withchasmophytic vegetation

20 18.2

AcknowledgmentsThe NSUH is commissioned and funded byNational Parks and Wildlife Service, Departmentof Environment, Heritage and Local Government.

ReferencesFossitt, J.A. (2000) Guide to Habitats in Ireland.The Heritage Council, Kilkenny

Perrin, P.M., Barron, S.J., Roche, J.R. &O’Hanrahan, B. (2010) Guidelines for a nationalsurvey and conservation assessment of uplandvegetation and habitats in Ireland. Version 1. IrishWildlife Manuals No. 48. National Parks andWildlife Service, Department of Environment,Heritage and Local Government, Dublin, Ireland.www.npws.ie/en/PublicationsLiterature/IrishWildlifeManuals

Roche, J.R. & Perrin, P.M. (in press) A newcounty record for alpine clubmoss (Diphasiastrumalpinum) from the Comeragh Mountains, Co.Waterford (H6). Irish Naturalists’ Journal.

Floral resources for bumblebees on Irishfarmland

V. Santorum and J. BreenDept. of Life Science, University of LimerickEmail: veronica.santorum@ul.ie

IntroductionWorldwide declines in bumblebees are echoed bylosses in Ireland. These are attributed mainly to theloss of flower-rich habitats (Fitzpatrick et al.,2007). Bee genetics and ecology, as well as ourhorticultural dependency upon pollinators, requiresthat bees are conserved over large geographicalareas. Bumblebee conservation is dependent uponmeasures applied on this type of broad scale.

Although the decline of bees is well established, todate, no measures directed at conserving‘pollinator diversity’ have been developed as partof Irish agri-environment schemes. In the UK,initiatives that target bees have been implementedwithin the Environmental Stewardship scheme andfound to be successful in supporting bees (Carvell2006). Providing flowers for forage has been acornerstone of the UK approach.

This brief study considers how abundant flowersare on Irish farms; identifies the major floralresources used by Irish bumblebees throughout thesummer and examines whether farms with moreflowers do support more bees. Bringing thisinformation together suggests approaches to beeconservation on Irish farmland. The aim is toinitiate debate and research that will stimulateaction to reverse bee declines in Ireland.

Materials and MethodsThe data presented is derived from three fieldstudies, carried out from 2003 to 2005 as part ofthe Ag-Biota project (Purvis et al., 2009). Allfarms were located in the eastern half of Ireland.

(a) Flower abundance on farmsThe abundance of all flowers, per square metre, inswards and as a percent of hedgerow length wasquantified during a study of 50 farms (2005).Composite flowers, umbels and other flowerclusters e.g. clover were each counted as oneflower. Farms were surveyed in early summer(May-June), mid-summer (July), late summer(August).

(b) Major floral resources for bumblebeesThe most frequently visited plant species wereidentified from observations of bees foragingduring transect surveys on 19 farms throughout thesummer (2003). (Period 1= 6 Jun – 14 Jul (42transects), Period 2 =15 Jul – 12 Aug (44 transects)

and Period 3= 14 Aug – 9 Sep (42 transects)respectively).

(c) Relationship between flower and beeabundancesThe relationship between abundances of bees andflowers was analysed using data from a pan-trapping survey of 18 farms (2004). Thepercentage cover of flowers was recorded using theBraun Blanquet cover scale and converted tomedian percentage cover values for analysis. Anegative binomial GLM was used to analyse therelationship between floral abundance and beeabundance.

Results and Discussion(a) Flower abundance on farmsAt each sampling period, 50% of swards had fewerthan 1 flower per 2 square metres (score of 1 onscale) (Figure 1). Over half of the hedges sampledhad flowers along at least a fifth of their length inevery sampling period.

1 2 3

0.0

0.2

0.4

0.6

0.8

1.0

Hedgerow

Period of summer

Flow

erA

bund

ance

(%he

dger

owle

ngth

)

1 2 3

01

23

45

Grassland

Period of summer

Flow

erA

bund

ance

(5po

ints

cale

)

Fig. 1. Boxplots of flower abundance in hedgerowsand grassland. (5 point scale used in grasslands:flowers absent=0; <0.5 flowers per m2=1; 0.5-1 flowerper m2= 2; 1-5 flowers per m2=3; 5-10 flowers perm2=4; >10 flowers per m2=5.)

(b) Major floral resources for bumblebeesOf a total of 684 observed foraging events, 84% offlower visits were to only five plant species (Table1). These flowers tended to have long floweringperiods.

Table 1. The plant species most visited by bees.(Early summer, E; mid+late summer, M+S.)Species % Periods in top 5Rubus fruticosus agg. 40.9 E, M+STrifolium repens 15.9 E, M+SCirsium vulgare 9.1 M+SCirsium arvense 9.0 E, M+SVicia sepium 7.0 E

(c) Relationship between flower and beeabundancesIn a sample of 407 bees, there was a significantrelationship between the abundance of bees and theabundance of flowers in the hedgerow (Fig. 2;negative binomial θ=6.62: n=16, F=24.2, p=0.0002, pseudo-R2=0.63) and a weak effect inresponse to flower abundance in the sward(negative binomial θ=2.62: n=16, F=5.2, p=0.04, pseudo-R2=0.27).

0 10 20 30 40

1020

3040

5060

Grassland

Flower Abundance

Bee

Abu

ndan

ce

0 20 40 60 80

1020

3040

5060

Hedgerow

Flower Abundance

Bee

Abu

ndan

ce

Fig. 2. GLM regression of bee abundance on thepercentage cover of flowers in grassland andhedgerow.

DiscussionHigher abundances of flowers in grasslands orhedgerows support larger numbers of bees. Fewagricultural grasslands or hedgerows, in the easternhalf of the country, have high abundances offlowers or bees. Flower abundances could beenhanced to boost bee numbers. Possible targets,based on this scant dataset, might be ~20% flowercover in grasslands and ~40% cover in hedgerows.Obviously, further study is required in the settingof such targets but the suggestion is made tostimulate discussion. Would targets be desirable?How could they be implemented and maintainedon the farm with ease and minimum costs? Oneapproach could be to plant flowers for forage.Another would be to encourage, via slight shifts inmanagement, more flowers from herbs, trees andshrubs already growing on farms. These shouldperhaps be regarded as complementary methodsrather than alternatives. Specifications in the Agri-Environment Options Scheme (AEOS), such as‘Traditional Hay Meadow’ and ‘Species richgrasslands’ offer opportunities to enhanceflowering within existing semi natural habitats.The alternative suggestion, planting forage forbees, is considered further in this paper.

Of the small range of flowers that are presentlysupporting the majority of bumblebees, all wereshown to be common species, most with longflowering periods. They are also perennials. Thesecharacteristics can be used to design a relativelycheap forage seed mix for bees that containsAsteraceae (daisy family) such as Centaurea as analternative to thistles and Fabaceae (legumefamily) such as clovers and vetches for the longertongued bees. Ideally the mix would flowercontinuously throughout the summer. Irish farmersare already planting for birds with the Wild BirdCover specification in AEOSS. Perhaps forage forbees could be integrated into this scheme. Howeveruptake of similar schemes has not been widespreadin the UK (Blake, 2011). A further option, thatrequires study, could be the incorporation of more‘bee-friendly’ legumes and a small number ofother herbs, into rotational grassland leys. If plantspecies were selected on the basis of nutritionalbenefits for livestock and bees as well as ease ofmanagement for persistence and flowering, uptakemight be on a much wider scale than wildlifespecific initiatives. If successful this couldintroduce flower-rich grasslands into intensivepastoral farming.

ConclusionsIncreasing the abundance of flowers on Irish farmswould have benefits for bees. This could beachieved through a combination of themanagement of existing habitats and new planting.Research is necessary to develop low-cost methodsthat are attractive to farmers as well as bees so thatthey are taken up on a large scale.

AcknowledgmentsEPA STRIVE Ag-Biota project for funding; and toland owners for access.

ReferencesBlake, R.J., Westbury, D.B., Woodcock, B.A.,Sutton, P. and Potts, S.G. (2011) Enhancing habitatto help the plight of the bumblebee. Pest ManagSci. 67:377-379

Carvell, C., Meek, W.R, Pywell, R.F., Goulson, D.and Nowakowski, M. (2006) Comparing theefficacy of agri-environment schemes to enhancebumble bee abundance and diversity on arablefield margins. J. App. Ecol. 44: 29-40

Fitzpatrick, U., Murray, T.E., Paxton, R.J., Breen,J., Cotton, D., Santorum, V. and Brown, M. (2007)Rarity and decline in bumblebees – A test ofcauses and correlated in the Irish fauna. Biol. Cons.136: 185-194.

Purvis, G. et al. (2009) Ag-Biota. STRIVE ReportSeries, Number 21. Johnstown Castle,Environmental Protection Agency.

Agri-environment management of grasslandfield margins

S.E. Spratt, A. Cooper and T. McCannEnvironmental Sciences Research Institute,University of Ulster, Coleraine, Northern Ireland,BT52 1SAEmail: Spratt-S1@email.ulster.ac.uk

IntroductionAgricultural grassland field margins associatedwith field boundary complexes such as hedges,earth banks and ditches have a distinctive plantcommunity structure (Marshall and Moonen, 2002)which can provide a partial reserve for plantspecies (Smart et al., 2006) in intensively managedlandscapes. There is a large research literature onthe community structure of arable field margins,but not for grass field margins.In Northern Ireland (NI), enclosed improved andsemi-improved grassland, described as ImprovedGrassland and Neutral Grassland in the UK BroadHabitats classification, cover an estimated 804,126hectares (56.8%) of NI. Field margin managementof these grassland types is an option of agri-environment schemes in NI. Improved Grassland isdominated by Lolium spp., whereas NeutralGrassland is managed less intensively and is morespecies rich. Our paper assesses the regionalspecies composition of Neutral Grassland fieldmargins to determine if their plant speciesassemblages are of particular biodiversity interestto the agriculture industry.

Materials and MethodsFrom the UK Neutral Grassland Broad Habitat, theNorthern Ireland Countryside Survey (NICS)Primary Habitat types A11 and A09 were selectedfor study (McCann et al., 2009). Theirmanagement intensity and species richness areintermediate to Improved Grassland and Semi-natural grassland Broad habitats.An area-proportional random sample of field-mapped A11 and A09 grassland patches from ¼kilometre sample grid squares was selected fromacross NI. Sample stratification was bymultivariate land class, the sampling rate was 1quadrat (4m2) per 5 hectares. A total of 212quadrats were surveyed comprising 106 fieldmargin quadrats (1m x 4m) and 106 pairedgrassland field quadrats (2m x 2m). Fieldwork wascarried out between late June and early September2009. Within each field, a 2m x 2m quadrat wasrandomly located within a random 100m2 (10m x10m) sub-sample. The field margin quadrat (1m x4m) was randomly located next to the nearest fieldboundary. In each quadrat, the percentage cover ofplant species was estimated according to theDOMIN scale.

Multivariate analysisField margin quadrats were classified by two-wayindicator species analysis using TWINSPAN forWindows 2.3 (Hill and Šmilauer, 2005).Pseudospecies cut levels were set as 0-5incorporating DOMIN values recorded. Theclassification was stopped by inspection to give 4groups. Multivariate ordination was carried outusing Canoco for Windows 4.5 (after Braak andŠmilauer, 2002). Detrended CorrespondenceAnalysis (DCA) was used to assess the maingradients of species composition. CanonicalCorrespondence Analysis (CCA) was used todetermine whether species where significantlymore abundant in the field margin quadratscompared with the grass field quadrats via MonteCarlo Permutation testing.

Results and DiscussionInspection of species lists of the field margin andgrass field quadrats gave a combined total of 164species (excluding trees and shrubs). There were105 species in the grass field quadrats and 150species in the field margin quadrats. Aclassification of the field margin quadrats gavefour end groups (Table 1).

Table 1. TWINSPAN classification groups.Vegetation group1 2 3 4

Mean no. spp. per quadrat 17 14 10 10Number of quadrats 13 55 18 22Proportion of sample (%) 13 50 17 20

The groups lay along the first axis of a DCAordination (Fig. 1) in the order 1, 2, 3 and 4, withgroups 1 and 2 separated by the second axis.Ellenberg soil nitrogen and pH indicator valuesand the Grime competition values showed a strongpositive correlation with the first axis of theordination. Grime stress tolerance values showed astrong negative correlation with axis 1. The mainordination axis, therefore, represents an inferredgradient of increasing soil nutrient status. Highvalues of the Ellenberg soil moisture indicatorspecies are correlated with axis 2. This axis,therefore, separates samples along a soil wetnessgradient.

Groups 1 and 2 of the field margin classification,characterised by stress-tolerant species of lownutrient status soils, had the highest biodiversity.Group 1 field margins were associated with speciesof better-drained soils, such as Festuca rubra andCynosurus cristatus.

Figure 1. First and second axes of a DCAordination showing groups of quadrats delimitedby a TWINSPAN classification along with specieswith a frequency of >75% in any one group and allTWINSPAN indicator species. Variables includedare; diversity (mean number of species perquadrat), Ellenberg indicator values for moisture,nitrogen, pH and light and Grime competitive,stress-tolerant and ruderal life strategies. Scalingfactor of explanatory variables = 7.25.Note: TWINSPAN groups are labelled as follows:Group 1 (○), Group 2 (X), Group 3 (+) and Group 4 (●). Axis 1 eigenvalue = 0.320, Axis 2 eigenvalue = 0.214 and Axis 3 eignevalue = 0.161.

Group 2 field margins were associated with speciesof poorly-drained soils, such as Juncus effusus.The area-proportional sampling system used meansthat the frequency of field margin quadrats in eachclass is directly proportional to their occurrence inthe countryside. Therefore, we estimate that group1 represents 13% of field margins and that group 2represents 50% of field margins. These twogroups, in particular group 1, have the greatestpotential value for targeting agri-environmentprescriptions. The remaining 47% of field margins(groups 3 and 4) have a relatively low plant speciesdiversity, not greatly different from the managedgrass field.

According to a CCA ordination, field marginquadrats have a significantly different speciescomposition compared with the grass fieldquadrats (F = 2.71) (p = 0.002). A t-value biplotshowed that there were 57 species more abundantin the field margin quadrats of which 12 weresignificantly more abundant. They were Cirsiumvulgare, Galium aparine, Lathyrus pratensis,Rubus fruticosus, Urtica dioica, Veronicachamaedrys, Vicia cracca and Vicia sepium(mainly scramblers and tall herbs of open habitats),Geranium robertianum, Hedera helix, Primulavulgaris and Viola riviniana (mainly broadleaf

woodland species). There were 41 species moreabundant in the grass field quadrats, six of whichwere significantly more abundant. These wereCardamine flexuosa/hirsuta, Carex panicea,Lolium perenne, Ranunculus flammula,Ranunculus repens and Trifolium repens, allcommon species of agricultural grasslands.

ConclusionsWe show that across NI as a whole, NeutralGrassland field margins, whether grazed or cut fora grass crop, have species composition different tothat of the managed part of the field and arecharacterised by species covering a wide range ofenvironmental tolerances and life strategies. Fieldmargin composition was determined largely by asoil nutrient gradient, from low to high nutrientstatus, in particular soil nitrogen. We conclude thatthe species diversity of field margins is directlyrelated to management inputs (in particularfertilizers) with more species-rich marginsassociated with a lower soil nutrient status andwith plant species considered to be agriculturalweeds associated with a higher soil nutrient status.Agri-environment management for increased plantspecies diversity in grassland field margins shouldaim to reduce nutrient inputs within 2m of fieldboundaries. Preferential funding, targeted atgrasslands with species-rich field boundaries suchas hedges and earth banks, from which dispersalinto the grassland margin could take place, wouldgive greater biodiversity gains.

AcknowledgmentsWe thank the University of Ulster field surveyorswho assisted with vegetation sampling. DARDfunded the research through a postgraduatestudentship.

ReferencesHill, M. O. and Šmilauer, P. (2005) TWINSPANfor Windows version 2.3. CEH and University ofSouth Bohemia, Huntingdom and CeskeBudejovice.

McCann, T., Rogers, D. and Cooper, A. (2009)Northern Ireland Countryside Survey 2007: FieldMethods and Technical Manual. NIEA, Belfast.

Smart, S. M., Marrs, R. H., Le Duc, M. G.,Thompson, K., Bunce, R. G. H., Firbank, L. G. andRossall, J. (2006) Spatial relationships betweenintensive land cover and residual plant speciesdiversity in temperate farmed landscapes. J ApplEcol. 43: 1128-1137.

ter Braak, C. J. F. and Šmilauer, P. (2002) Canocofor Windows version 4.5. Biometris-Plant ResearchInternational,Wageningen, The Netherlands.

Mapping High Nature Value farmland inIreland

C.A. Sullivan1, D.A. Bourke2, J.A. Finn3, M.Sheehy Skeffington4 and M. J. Gormally5

1Humboldt State University, Arcata, CA, 95521,USA 2 Trinity College Dublin, 3TeagascEnvironmental Research Centre, Johnstown Castle,Co. Wexford, 4Botany and Plant Science, NUIGalway 5 Applied Ecology Unit, NUI Galway.Email: caroline.sullivan1@gmail.com

IntroductionFarming sustains the high nature value of manyagricultural landscapes across Europe and isintegral to maintaining their biodiversity (Bignaland McCracken, 1996). High Nature Value (HNV)farmland occurs where agriculture sustains a highspecies and habitat diversity (Anderson et al.,2004). The identification of High Nature Value(HNV) farmland throughout Europe has become animportant goal for EU Member States. They arerequired to ensure that their current RDPs (2007-2013) put priority on HNV farmland identification,support and maintenance and to monitor anychanges in HNV farmland extent (CEC, 2006).However, the range of HNV farmland and farmingsystems throughout the EU leads to challenges foreach member state in identifying these areas.Existing nature conservation designations will, atbest, protect a minority of HNV farmland and donot target areas of high farmland biodiversitywithin more intensively-managed agriculturallandscapes (Henle et al., 2008). Spatial targeting ofHNV farmland is the key to providing the supportnecessary to maintain these HNV landscapes,particularly those outside of designated sites(Sullivan et al, 2011). This paper examines the useof modelling to map areas of HNV farmland. Itfocuses on modelling semi-natural habitats inparticular as, based on the current definitions ofHNV farmland, identification of HNV reliesheavily on semi-natural habitat cover. Ifsuccessful, this model could have many otheruseful applications for planning and localgovernment.

Materials and MethodsThirty two farms were selected randomly from sixdifferent District Electoral Divisions (DEDs) ineast Galway outside of designated areas. Allhabitats on each farm were identified according toFossitt (2000). Farm management data such asstocking density, farming enterprise and reseedingpractices were collected from each farmer at thetime of field sampling. The habitat data werecollected from May to October of 2006 and 2007.All farms and habitat areas were digitised andareas calculated using ArcGIS 9.3. Fields and

hedgerows were surveyed in further detail. A rangeof variables that might explain the habitat diversityon the farms was selected based on a literaturereview (Table 1). Both farm-level and landscape-level variables were considered. Factors affectingsemi-natural habitat cover were modelled usingGeneral Additive Modelling (GAM). See Sullivanet al. (2011) for further details.

Table 1. Explanatory variables tested in GAMmodel to predict semi-natural habitat area onfarms. * indicates distance to nearest feature.

VariableField boundary density Secondary road*Field boundary length 3rd class road*Farm enterprise 4th class road*DED Any class of road*AE scheme participation City*Average field size Town*Stocking density (LU/ha) Lake*Soil diversity index SAC*Elevation SPA*River and stream length NHA*Regional road* pNHA*Primary road* Native woodland*

Results and DiscussionFarmland habitat diversityThe average cover of semi-natural habitats on afarm was 15.2% (±3.0 s.e.). This figure variedfrom 0% to just over 60% (Fig. 1), with just threefarms having no non-linear semi-natural habitatcover. All farms surveyed were grassland-dominated, and a total of 24 habitats were recordedon the 32 farms. More than 50% of those weresemi-natural habitats.

Fig. 1. Percentage semi-natural habitat per farm(by habitat type). Three farms with no semi-naturalhabitats are excluded.

Semi-natural grassland (predominantly WetGrassland) was a common component of the semi-natural habitat cover on the majority of those farmswith semi-natural habitats (See Fig. 2 for examplesfarm habitat maps). These data illustrate thevariation in semi-natural habitat diversity as wellas the spatial location of the habitats that occur onlowland farms.

Fig. 2. Habitat composition of a) a 15.6farm with 7% semi-natural habitat cover a25.3 ha beef farm with 47% semi-naturalcover.

Modelling semi-natural habitatsThe GAM analysis demonstrated that thesemi-natural habitat cover can be predictthe variables. There were five models whintercept and all variables gave statsignificant results. Soil diversity anddensity featured as explanatory variables toextent of semi-natural habitat in three ofmodels. Using this method, higher levelsnatural habitat cover are indicated bystocking density, higher soil diversity andriver and stream lengths.

Future applicationsThis model could be applied to all farmsGalway, providing a map of potential semihabitat areas in the region. Further ground-would allow accuracy testing and refinememodel.This modelling approach could be modifiedin different regions and different countries.could be adapted to help identify areas thatthreshold levels of semi-natural habitat coadditional criteria), thereby indicating thpresence of HNV farmland. If an area o

meets these criteria, then they could be targeted formore detailed investigation of their HNV status(e.g. Sullivan et al., 2010).Modelling the spatial distribution of semi-naturalhabitat area would also lead to improved spatialtargeting of management actions for natureconservation by local councils. National maps ofsemi-natural habitat cover on farms would alsobenefit agri-environment policy decisions as areasidentified with low or no semi-natural habitatcover could be targeted for biodiversity restorationprojects or recruitment of farmers in certain low-biodiversity regions into agri-environment

b

schemes.

Acknowledgments

a

ha beefnd b) ahabitat

percented fromere the

isticallystocking

predictthe fiveof semi-

lowerlonger

in east-naturaltruthingnt of the

for useThus, itsurpass

ver (ande likelyr region

This work was funded by Teagasc under the WalshFellowship Scheme.

ReferencesAnderson et al. (2003) Developing a High NatureValue farming area indicator. EEA, EuropeanEnvironment Agency, Copenhagen.

Bignal, E.M., McCracken, D.I. (1996) Low-intensity farming systems in the conservation ofthe countryside. J. Appl. Ecol. 33: 413-424.

CEC. (2006) Council Decision of 20 February2006 on Community strategic guidelines for ruraldevelopment (programming period 2007 to 2013).Official Journal of the European Union.

Fossitt, J., (2000) A guide to habitats in Ireland.The Heritage Council, Kilkenny.

Henle et al. (2008) Identifying and managing theconflicts between agriculture and biodiversityconservation in Europe- a review. Agric., Ecosys.& Env. 124: 60-71.

Sullivan, C.A., Skeffington, M.S., Gormally, M.J.,Finn, J.A. (2010) The ecological status ofgrasslands on lowland farmlands in western Irelandand implications for grassland classification andnature value assessment. Biol. Cons. 143: 1529-1539.

Sullivan, C.A., Bourke, D., Skeffington, M.S.,Finn, J.A., Green, S., Kelly, S., Gormally, M.J.(2011) Modelling semi-natural habitat area onlowland farms in western Ireland. Biol. Cons. 144:1089-1099.

Tematea. (2008) Kyiv Resolution on Biodiversity,http://www.tematea.org/?q=node/1349&PHPSESSID=d625dccb8dc33fa24b9df7254e923c9d.Tematea.

Impact of riparian vegetation structure onsmall mammal communities

D. Ó hUallacháin and D. MaddenTeagasc, Environmental Research Centre,Johnstown Castle, Co. WexfordEmail: daire.ohuallachain@teagasc.ie

IntroductionIntensification of agriculture over the last numberof decades has led to a dramatic change inagricultural production methods. This in turn hasresulted in a loss of ecological heterogeneity (Petitand Firbank, 2006) and has contributed to the lossof biodiversity (Robinson and Sutherland, 2002),resulting in significant implications for wildspecies of flora and fauna.Protection of uncultivated field margins,hedgerows, ditches and watercourse margins isvital to ecosystems, because they are anincreasingly important source of seed andinvertebrate food (Wilson et al, 1999). Smallmammals on farmland are largely confined tothese areas of non-crop habitat and are thereforeparticularly vulnerable to agriculturalintensification (Bates and Harris, 2009).The Rural Environment Protection Scheme(REPS) guidelines (Measure 3) require farmers tofence all water-course margins to prevent bovineaccess. No further management of these fencedareas is required. Such fencing will give rise tonatural succession of vegetation from grassy(primarily herbaceous species) to scrubby (e.g.bramble, gorse) to woody (e.g. alder, willow)vegetation. These fenced sites are seen aspotential refuges for biodiversity, however, littleinformation exists in relation to the effect thatsuccession of the margin vegetation has on smallmammal communities. Small mammals constitutethe main prey biomass that influences thediversity and number of predator species such askestrel Falco tinnunculus, barn owl Tyto alba(Red-listed species) and pine marten Martesmartes, thus, contributing to the complexity offood webs (Korpimaki and Norrdahl, 1991).The aim of this study was to highlight the impactof vegetation succession in watercourse marginson small mammal communities.

Materials and MethodsThe study took place on a number of sites in Co.Wexford. Thirty-metre stretches of watercoursemargin were selected for study. All watercourseswere between 1 m and 3 m in width and flowedfor at least nine months of the year. A total of 42sites were selected (14 grassy, 14 scrubby and 14woody). Each site was dominated by either grassyvegetation, scrubby vegetation or woodyvegetation. Grassy sites were dominated by

gramineous plants such as Lolium and Agrostisand by forbs such as Cirsium arvense andmembers of the Ranunculaceae, Polygonaceae,and Leguminosae families. Scrubby sites(vegetation less than 2 metres in height) weredominated by Ulex europaeus, Rubus fruticosus,Prunus spinosa and members of the Umbelliferaefamily. Woody sites (vegetation above 2 metres inheight) were dominated by Crataegus monogyna,Fraxinus excelsior and by members of familySalicaceaeSmall mammals at sites were sampled usingLongworth traps. Traps were placed in pairs every10m in trap lines, at distances of 1 m and 5 mfrom the stream edge. Therefore, each 30 msection of margin contained 16 traps. Traps wereleft in situ for two nights during each samplingsession and inspected at dawn, dusk and at leastonce during the day. Captured mammals weremarked and weighed, with additional informationsuch as sex, age group, breeding condition andlength of hind foot being noted. The results in thispaper are based on three sampling sessions, earlysummer (May), late summer (August) and winter(December) in 2007.Abundance data were analysed using SAS and asplit plot in time analysis with Poissondistribution and log-transformed data. TheShannon-Weiner Index was used when assessingspecies diversity of each habitat. The weights ofsmall mammals were analysed using ANOVA.

Results and DiscussionThe results in this paper are gleaned from 4,032trap nights in 2007. A total of 317 capturesoccurred, of these, 90.5% were woodmouse(Apodemus sylvaticus), 7.6% were shrew (Sorexminutes) and 1.9% were house mouse (Musdomesticus). The remaining small mammalspecies found in Ireland, the bank voleClethrionomys glareolus and greater white-toothed shrew Crocidura russula are not found inthe study area.From the results, it was found that habitat had asignificant effect (P < 0.01) on small mammalabundance. Significantly more mammals werecaught in woody habitats as opposed to scrubby orgrassy habitats. Scrubby habitats had the lowestcapture rate (the lower rate for grassy habitatsapparent in Table 1 is as a result of a high numberof traps catching non-target species in these sites).The low capture rate in scrubby habitats is largelydue to the availability of food. Habitats such asgorse, despite providing excellent cover for smallmammals, provide considerably less of theirfavoured food than many woody sites(Montgomery et al., 1991). This theory is givenfurther credence by the fact that woodmousefound in scrub, weighed significantly less than

those found in either grassy or woody habitats (P< 0.05).More mammals were caught at a distance of 1metre as opposed to 5 metres from the streamedge (P< 0.0001).Sampling period (early summer, late summer andwinter) also played a significant role (P < 0.05)with more small mammals being caught in winterthan in either early or late summer. This is largelydue to the fact that small mammals are lessterritorial in winter and must also travel greaterdistances to forage, therefore increasing theirlikelihood of encountering a trap.

In relation to the condition of small mammals(woodmouse), distance from stream, season, sexand breeding condition had a significant effect onthe weight of adult woodmice.Mice trapped adjacent to the stream weighedsignificantly more than those trapped at a distanceof 5m from the stream (P < 0.05). Males weresignificantly heavier than females (P < 0.001) andbreeding mice were significantly heavier thannon-breeding mice (P < 0.001). This latter point isnot surprising considering that during the breedingseason, the genitals swell to many times their non-breeding weight (Gurnell and Flowerdew, 2006)thus increasing the overall weight. There was asignificant correlation (P < 0.001) between theweight and the tarsus length of the animal. Bothmeasurements are used as an indicator of fitness.

Although abundance of small mammals is greatestin woody habitats, these sites were the leastdiverse (Shannon-Weiner index in Table 1).Grassy habitats showed the most diversity withthe small mammal community consisting of 78%woodmouse, 19% pygmy shrew and 3% housemouse, whereas the community of woody habitatswere dominated (98%) by woodmice.

Table 1. Small mammal captures in watercoursemargins (aggregated across 14 replicate sites overthree seasonal sampling events).

ConclusionsSmall mammals play an important role inagricultural ecosystems, and managementprescriptions that promote and enhance theirabundance and diversity should be promoted.

Increased small mammal populations on farmlandare crucial to improving the biodiversity ofagricultural ecosystems (Bates and Harris, 2009).The current REPS guidelines of fencing allwatercourse margins, and the subsequentsuccession it gives rise to, is not promoting smallmammal diversity within riparian habitats.Watercourse margin management prescriptionswhich include fencing (and the resultantsuccession of vegetation) do not provide a suitablehabitat for pygmy shrews (Bern Convention,Annex II species).It is likely that periodical cutting, grazing andalternative managements of margins wouldpromote heterogeneity and in-turn give rise to agreater diversity of small mammals and also offloral and faunal communities in general.

AcknowledgmentsThis project was funded by the Department ofAgriculture, Fisheries and Food under theNational Development Plan.

ReferencesBates, F.S. & Harris, S. (2009). Does hedgerowmanagement on organic farms benefit smallmammal populations? Agric. Ecosyst. andEnviron. 129: 124-130.

Gurnell, J. & Flowerdew, J. R. (2006). LiveTrapping Small Mammals. The Mammal Society.

Korpimaki, E. & Norrdahl, K. (1991). Numericaland functional responses of kestrels, short-earedowls, and long-eared owls to vole densities. Ecol.72: 814-826

Montgomery, W.I. & Dowie, M. (1993). Thedistribution of the wood mouse Apodemussylvaticus and the house mouse Mus domesticuson farmland in north-eastern Ireland. Irl Nat J. 24:199-203.

Petit, S. & Firbank, L. (2006). Predicting the riskof losing parcels of semi-natural habitat tointensive agriculture. Agric. Ecosyst. & Environ15: 277-280.

Robinson, R.A. & Sutherland, W.J. (2002). Post-war changes in arable farming and biodiversity inGreat Britain. J. App. Ecol. 39: 157-176.

Wilson, J.D., Morris, A.J., Arroyo, B.E., Clark,S.C. & Bradbury, R.B. (1999). A review of theabundance and diversity of invertebrate and plantfoods of granivorous birds in northern Europe inrelation to agricultural change. Agric. Ecosyst. &Environ. 75: 13-30

Grassy Scrubby WoodyNo. caught at 1 m 63 54 108No. caught at 5 m 11 54 27Total capture 74 108 135Shannon-WeinerIndex (H’)

0.26 0.17 0.05

Adult woodmousemean weight (g)

23.57 21.72 22.64

Adult Woodmousemean tarsus (mm)

13.26 12.96 13.04

Acknowledgements

We gratefully acknowledge financial support for this conference from the Department ofAgriculture, Food and Marine through the Research Stimulus Fund under the NationalDevelopment Plan 2006. This was provided through project RSF 06 382 ‘An evaluation ofexisting and potential measures to sustain an increased biodiversity and water quality on Irishfarms’.

We are especially grateful to Mary Foley and Aileen Cardiff for their administrative supportin the weeks and months in advance of the conference. Alison Maloney assisted with thepreparation of the conference proceedings and flier. We also thank the following for theirassistance: Susan Carr, Owen Carton, Eric Donald and Mairead Esmonde.

We thank our keynote speakers, and contributors of oral and poster presentations for sharingtheir information and insights. Finally, we thank the attendees for their engagement andcontributions.