Assessing plant diversity and composition in grasslands across spatial scales: the standardised EDGG sampling methodology

Abstract

This paper presents the details of the EDGG sampling methodology and its underlying rationales. The methodology has been applied during EDGG Research Expeditions and EDGG Field Workshops since 2009, and has been subsequently adopted by various other researchers. The core of the sampling are the EDGG Biodiversity Plots, which are 100‐m² squares comprising, in two opposite corners, nested‐plot series of 0.0001, 0.001, 0.01, 0.1, 1 and 10 m² square plots, in which all terricolous vascular plants, bryophytes and lichens are recorded using the shoot presence method. In the 10‐m² plots, species cover is also estimated as a percentage and various environmental and structural parameters are recorded. Usually the EDGG Biodiversity Plots are complemented by the sampling of additional 10 m² normal plots with the same parameters as the 10‐m² corners of the first, allowing coverage of a greater environmental diversity and the achievement of higher statistical power in the subsequent analyses for this important grain size. The EDGG sampling methodology has been refined over the years, while its core has turned out to generate high‐quality, standardised data in an effective manner, which facilitates a multitude of analyses. In this paper we provide the current versions of our guidelines, field forms and data entry spreadsheets, as open‐access Online Resources to facilitate the easy implementation of this methodology by other researchers. We also discuss potential future additions and modifications to the approach, among which the most promising are the use of stratified‐random methods to a priori localise the plots and ideas to sample invertebrate taxa on the same plots and grain sizes, such as grasshoppers (Orthoptera) and vegetation‐dwelling spiders (Araneae). As with any other method, the EDGG sampling methodology is not ideal for every single purpose, but with its continuous improvements and its flexibility, it is a good multipurpose approach. A particularly advantageous element, lacking in most other sampling schemes, including classical phytosociogical sampling, is the multi‐scale and multi‐taxon approach, which provides data that allow for deeper understanding of the generalities and idiosyncrasies of biodiversity patterns and their underlying drivers across scales and taxa.

Full text

Assessingplantdiversityandcompositioningrass‐
landsacrossspatialscales:thestandardisedEDGG
samplingmethodology
JürgenDengler*,1,2,SteffenBoch3,GoffredoFilibeck4,AlessandroChiarucci5,IwonaDembicz6,1,RiccardoGuarino7,
BenjaminHenneberg8,MonikaJanišová9,CorradoMarcenò10,AlirezaNaqinezhad11,NinaY.Polchaninova12,Kiril
Vassilev13&IdoiaBiurrun14
1)PlantEcology,BayreuthCenterofEcologyandEnvironmentalResearch
(BayCEER),UniversityofBayreuth,Universitätsstr.30,95447Bayreuth,GER‐
MANY;juergen.dengler@uni‐bayreuth.de
2)GermanCentreforIntegrativeBiodiversityResearch(iDiv)Halle‐Jena‐
Leipzig,DeutscherPlatz5e,04103Leipzig,GERMANY
3)InstituteofPlantSciences,UniversityofBern,Altenbergrain21,3013Bern,
SWITZERLAND;steffen.boch@ips.unibe.ch
4)DepartmentofAgriculturalandForestrySciences,UniversityofTuscia,
01100Viterbo,ITALY;filibeck@unitus.it
5)DepartmentofBiological,GeologicalandEnvironmentalSciences,Alma
MaterStudiorumUniversityofBologna,ViaIrnerio42,40126Bologna,ITALY;
alessandro.chiarucci@unibo.it
6)DepartmentofPlantEcologyandEnvironmentalConservation,Facultyof
Biology,BiologicalandChemicalResearchCentre,UniversityofWarsaw,ul.
ŻwirkiiWigury101,02‐089Warsaw,POLAND;iwodem@op.pl
7)Dept.STEBICEF−BotanicalUnit,UniversityofPalermo,viaArchirafi38,
90123Palermo,ITALY;guarinotro@hotmail.com
8)AnimalEcologyII,BayreuthCenterofEcologyandEnvironmentalResearch
(BayCEER),UniversityofBayreuth,Universitätsstr.30,95447Bayreuth,GER‐
MANY;ben_henneberg@web.de
9)InstituteofBotany,SlovakAcademyofSciences,Ďumbierska1,97400
Banská‐Bystrica,SLOVAKIA;monika.janisova@gmail.com
10)DepartmentofBotanyandZoology,FacultyofScience,MasarykUniver‐
sity,Kotlářská2,61137Brno,CZECHREPUBLIC;marcenocorrado@libero.it
11)DepartmentofBiology,FacultyofBasicSciences,UniversityofMazanda‐
ran,P.O.Box47416‐95447,Babolsar,Mazandaran,IRAN;
a.naqinezhad@umz.ac.ir
12)V.N.KarazinKharkivNationalUniversity,4SvobodySq.,61077Kharkiv,
UKRAINE;polchaninova_n@ukr.net
13)InstituteofBiodiversityandEcosystemResearch,BulgarianAcademyof
Sciences,23Acad.G.BonchevStr.,1113Sofia,BULGARIA;kiril5914@abv.bg
14)DepartmentofPlantBiologyandEcology,UniversityoftheBasqueCoun‐
tryUPV/EHU,PO.Box644,48080Bilbao,SPAIN;idoia.biurrun@ehu.es
*)Correspondingauthor
BulletinoftheEurasianGrasslandGroup32(2016):13‐30
Abstract:ThispaperpresentsthedetailsoftheEDGGsamplingmethodologyanditsunderlyingrationales.Themethodologyhasbeen
appliedduringEDGGResearchExpeditionsandEDGGFieldWorkshopssince2009,andhasbeensubsequentlyadoptedbyvarious
otherresearchers.ThecoreofthesamplingaretheEDGGBiodiversityPlots,whichare100‐m²squarescomprising,intwoopposite
corners,nested‐plotseriesof0.0001,0.001,0.01,0.1,1and10m²squareplots,inwhichallterricolousvascularplants,bryophytes
andlichensarerecordedusingtheshootpresencemethod.Inthe10‐m²plots,speciescoverisalsoestimatedasapercentageand
variousenvironmentalandstructuralparametersarerecorded.UsuallytheEDGGBiodiversityPlotsarecomplementedbythe
samplingofadditional10m²normalplotswiththesameparametersasthe10‐m²cornersofthefirst,allowingcoverageofagreater
environmentaldiversityandtheachievementofhigherstatisticalpowerinthesubsequentanalysesforthisimportantgrainsize.The
EDGGsamplingmethodologyhasbeenrefinedovertheyears,whileitscorehasturnedouttogeneratehigh‐quality,standardised
datainaneffectivemanner,whichfacilitatesamultitudeofanalyses.Inthispaperweprovidethecurrentversionsofourguidelines,
fieldformsanddataentryspreadsheets,asopen‐accessOnlineResourcestofacilitatetheeasyimplementationofthismethodology
byotherresearchers.Wealsodiscusspotentialfutureadditionsandmodificationstotheapproach,amongwhichthemostpromising
aretheuseofstratified‐randommethodstoapriorilocalisetheplotsandideastosampleinvertebratetaxaonthesameplotsand
grainsizes,suchasgrasshoppers(Orthoptera)andvegetation‐dwellingspiders(Araneae).Aswithanyothermethod,theEDGG
samplingmethodologyisnotidealforeverysinglepurpose,butwithitscontinuousimprovementsanditsflexibility,itisagoodmulti‐
purposeapproach.Aparticularlyadvantageouselement,lackinginmostothersamplingschemes,includingclassicalphytosociogical
sampling,isthemulti‐scaleandmulti‐taxonapproach,whichprovidesdatathatallowfordeeperunderstandingofthegeneralities
andidiosyncrasiesofbiodiversitypatternsandtheirunderlyingdriversacrossscalesandtaxa.
Keywords:biodiversity;bryophyte;EDGGBiodiversityPlot;invertebrate;lichen;methodology;multi‐taxonstudy;relevé;scale‐
dependence;speciesrichness;vegetation‐environmentrelationship;vegetationsampling.
Abbreviations:EDGG=EurasianDryGrasslandGroup;GIS=geographicinformationsystem;QA=qualityassessment;SAR=species‐
arearelationship..
ThisarticlecontainsOnlineResources,whichareavailablefromtheEDGGhomepage(http://www.edgg.org)aswellasfromthe
ResearchGateaccountofthefirstauthor(https://www.researchgate.net/profile/Juergen_Dengler).
Researchpaper
13 Bulletin of the Eurasian Dry Grassland Group 32 October 2016

Introduction
Understandingtheunequaldistributionofspeciesdiversity
isoneofthegreatestchallengesinecology.Standardized
samplingprotocolsfordiversityassessmentsaretherefore
essentialtoreflectdiversitypatternsacrossspatialscales
andtocomparethediversitiesofdifferentecosystems.
Palaearcticgrasslandsharbourahighdiversityofvarious
taxa(Allanetal.2014)andholdthemajorityofworldre‐
cordsinvascularplantspeciesrichnessforgrainsizes
smallerthan100m²(Wilsonetal.2012;Dengleretal.2014;
Chytrýetal.2015).Inaddition,bryophyteandlichendiver‐
sitycanalsobehighinthesehabitats(Dengler2005;Müller
etal.2014;Bochetal.2016;Dengleretal.2016).However,
therearealsoparticularlyspecies‐poorgrasslandtypesin
thePalaearctic(Dengler2005;Dengleretal.2016),making
Palaearcticgrasslandsasawholesuitableasamodelsystem
toanalysediversitypatternsandtheirunderlyingdrivers.
Theacquisitionofknowledgeonthesetopicsisofgreatim‐
portanceinthedevelopmentofappropriateconservation
measuresandinordertomaintainthesehighlydiverseeco‐
systemsandtheecosystemfunctionstheyprovide(Soliveres
etal.2016).
Themajorityofstudiesanalysingtheeffectsofabiotic,biotic
andhistoricalfactorsonspeciesdiversityimplicitlyassume
thatthesefactorsareuniversal,andthusstudyingbiodiver‐
sitypatternsatonegrainsizeprovidesanswersforallgrain
sizes.Onthebasisofthenowadaysreadilyavailableand
relativelystandardisedcoarse‐graindata,mostsuchstudies,
andthusgeneralecologicalknowledge,arebasedoncoarse‐
grainanalyses.Thesetypicallyrelyondatacollectedatgrain
sizesofhundredsorthousandsofsquarekilometres,while
fine‐grainanalysesacrosslargespatialextentsarelargely
lacking(Becketal.2012).However,ithaslongbeenhy‐
pothesizedthattheprevailingdriversofbiodiversityvary
stronglybetweengrainsizes(Shmida&Wilson1985).This
assumptionhasindeedfoundstrongsupportinseveralre‐
centmeta‐analyses(Fieldetal.2009;Siefertetal.2012).
Studyingpatternsanddriversofbiodiversityatsmallgrain
sizesoverseveralordersofmagnitudecanbeparticularly
insightful,asatthislevel,(plant)individualsofdifferentspe‐
ciesinteractwitheachotherandtheirenvironment(see
examplesinReedetal.1993;Dupré&Diekmann2001;de
Belloetal.2007;Giladietal.2011;Turtureanuetal.2014).
However,suchstudiesarestillrareandmainlyrestrictedto
thelocal,orveryrarelytotheregionalscale.Oftencompari‐
sonsofstudies,orevenjointanalysesoftheircombined
data,areimpededbytheidiosyncrasiesoftheplotsizesand
samplingschemesused.Thesituationisevenworseforphy‐
tosociologicaldatathatareavailableinlargequantitiesand
aresuitableformanypurposes(Dengleretal.2011;Chytrý
etal.2016),asplotsizes(Chytrý&Otýpková2003),aswell
assamplingquality(Chytrý2001),varygreatly.Thus,such
phytosociologicallegacydataareacomplexsourceforstud‐
iesondiversitypatternsandtheirscaledependence.
Bearingthisinmind,standardisedmulti‐scalediversitysam‐
plingschemes,oftencombinedwiththesamplingofabiotic
factorsandsometimesalsonon‐planttaxa,havebeenpro‐
posed,amongthemtheWhittakerplots(Shmida1984),the
plotsoftheCarolinaVegetationSurvey(CSW;Peetetal.
1998)andtheBIOTASouthObservatories(Jürgensetal.
2012).Inspiredbythese,aswellasbysimilarattemptsby
colleagues(Hobohm1998;Dolnik2003),studentsofthefirst
authortestedtheseideasintheirtheses(Löbel2002;Boch
2005;Allers2007).Onthebasisofthesestudies,Dengler
(2009)thenproposedtheso‐calledflexiblemulti‐scaleap‐
proachforstandardisedrecordingofplantspeciesrichness
patterns,whichcanbeseenasamethodologicalframework
thatallowsmanydifferentimplementations,butwitha
commoncore.Startinginthesameyear,thissamplingap‐
proachgaverisetotheResearchExpeditionsoftheEuro‐
peanDryGrasslandGroup(EDGG;http://www.edgg.org;
Vrahnakisetal.2013),whichweremeanwhilerenamedas
FieldWorkshopsoftheEurasianDryGrasslandGroup(Venn
etal.2016).Hereweusetheterm“fieldpulse”toreferto
bothtypes,inspiredbytheCarolinaVegetationSurvey(Peet
etal.1998),“pulse”implyinganintensiveeventofrelatively
shortduration,butrepeatedovertime.Thefirsteventin
Transylvaniain2009(Dengleretal.2009;Dengleretal.
2012a;Turtureanuetal.2014)wasfollowedbyeightmore
internationallyattendedfieldpulsesconductedfromSpain
inthewesttoSiberiaintheeastandfromSicilyinthesouth
toPolandinthenorth(Vrahnakisetal.2013;Vennetal.
2016).Thesefieldpulsescreatedahugecommondatapool
forjointanalyses(Dengleretal.2016)andyieldedawhole
seriesofpapersondiversitypatterns(Turtureanuetal.
2014;Kuzemkoetal.2016;Polyakovaetal.2016),andalso
onspeciescompositionandsyntaxonomy(Dengleretal.
2012a;Pedashenkoetal.2013;Kuzemkoetal.2014).While
thesamplingapproachgenerallyturnedouttobeveryeffec‐
tiveforawiderangeofdifferentresearchquestions,the
jointfieldworkalsoledtonumeroussmallmodificationsand
additions.Moreover,participantsinthefieldpulsesadopted
thesamplingmethodsintheirownprojects(e.g.Baumann
etal.2016;Cancellierietal.2017;M.J.andcolleaguesin
Ukraine,unpublished)andevenresearchersnotrelatedto
theEDGGstartedtousethisapproach(e.g.Mardari&
Tănase2016;A.C.andcolleaguesinItaly,unpublished).
TheEDGGsamplingapproach,withtheEDGGBiodiversity
Plotsasitscoreelement,isthusevidentlyeffectiveandat‐
tractive.Todate,however,nocompletein‐depthandup‐to‐
datedescriptionofthisapproachhasbeenpublished.Ac‐
cordinglythispaperpresentsthecurrentversionofourap‐
proach,withthelatestmodifications,subsequenttothe9th
EDGGFieldWorkshop2016inSerbia,criticallyassessingits
prosandconsaswellaspotentialextensionsanddemon‐
stratingpotentialapplications.Webelievethatourproposal
andrationalescanalsocontributetoabetterstandardisa‐
tionofothersamplingapproaches,forexample,inphytoso‐
ciology(compareMucinaetal.2000).Tofacilitatetheadop‐
tionormodificationofourapproachinotherstudies,we
providethesamplingformsandspreadsheetsfordatahan‐
dlinginconjunctionwiththisarticle.

14 Bulletin of the Eurasian Dry Grassland Group 32 October 2016

DescriptionoftheEDGGsamplingmethodology
anditsrationale

Thedescriptionofthemethodologyisalwaysindicatedin
bold‐italics,followedbythejustificationinnormalfont.The
outlinedmethodologyhasbeenappliedintheEDGGfield
pulsessince2009(Dengleretal.2009),unlessindicatedoth‐
erwise.Whereappropriate,themethodologicalexplanations
areconcludedwithpracticalhintsfortheirimplementation
initalics.

A.Locationandarrangementoftheplots
A.1Ineachstudysite,theEDGGBiodiversityPlots(100m²)
areselectedsubjectivelyinquasi‐homogenousstandsofad‐
hocrecognizabledifferentvegetationtypesregardingboth
siteconditionsandfloristiccomposition(Photos1−6).This
approachaimsatencompassingasmuchaspossiblethegeo‐
graphicandecologicalheterogeneitywithintheaprioride‐
fined“studyuniverse”(e.g.allwetgrasslandsofaregion).
Unlikethepracticeofsomephytosociologists(seeGlavac
1996),theoccurrenceofdiagnosticspeciesorconcurrence
withrecognisedsyntaxaareexplicitlyexcludedasselection
criteria.Ourapproachontheonehandensuresthatecologi‐
calgradientsarerepresentativelycoveredwithalimited
samplesize,i.e.spatiallyraretypesarerelativelyover‐
represented,whichisimportantforanalysesofdiversity‐
environmentrelationships.Ontheotherhand,limitingthe
numberofbiodiversityplotspersiteavoidstheriskofover‐
samplingandpseudo‐replication.Withtheimplicitphiloso‐
phyofrelatingthenumberofbiodiversityplotspersitetoits
ecologicalheterogeneity,ourapproachmimicsadhocthe
post‐hocheterogeneity‐constrainedrandomresampling
(Lengyeletal.2011).

A.2Thestudy‐plotsizesare1cm²;10cm²,100cm²,1000
cm²,1m²,10m²and100m²(Fig.1).Usingplotsizesalways
differingbyoneorderofmagnitudeisalsothephilosophyof
otherwidespreadmulti‐scaleapproaches.Forexample,
Shmida(1984),Peetetal.(1998)andJürgensetal.(2012)
usethesamesetofplotsizes,excludingonlythesmallest
onesandadding1000m².Theseplotsizesalsoincludethree
ofthemostfrequentlyusedplotsizesinphytosociology,
namely1,10and100m²(Chytrý&Otýpková2003).Having
theplotsizesonageometricscaleisbeneficialformanyana‐
Photo1.EDGGBiodiversityPlotduringtheEDGGFieldWork‐
shopinSicily,Italy,2012(Photo:T.Becker).
Photo2.EDGGBiodiversityPlotduringtheEDGGResearch
ExpeditioninKhakassia,Russia,2013(Photo:J.Dengler).
Photo3.EDGGBiodiversityPlotduringtheEDGGField
WorkshopinNavarre,Spain,2014(Photo:J.Dengler).
Photo4.EDGGBiodiversityPlotduringtheEDGGField
WorkshopinSerbia,2016(Photo:J.Dengler).
15 Bulletin of the Eurasian Dry Grassland Group 32 October 2016

lyticalpurposes,whilethetenfoldareaincreasefromone
plottothenextlargestoneislesssampling‐intensiveand
avoidsunusualsizes(like256m²),whichoccurinarea‐
doublingapproaches(e.g.Chiaruccietal.2006).Wedidnot
include1000m²inourstandardprocedurebecausecomplete
samplingofsuchanareainspecies‐richPalaearcticgrass‐
landscanbeextremelytime‐consuming.Forexample,Dolnik
(2003),whoisaveryexperiencedfieldbotanist,neededupto
sevenhourstosamplenestedplotsofupto900m²(without
replicationofsubplots)innotparticularlyrichgrasslandtypes
oftheCuronianSpit(Russia).Incontrast,addingsmallergrain
sizescomparedtotheotherstandardsamplingschemes,re‐
quiresonlyminimalextraeffortbutishighlybeneficialfor
analysessuchasspecies‐arearelationships(SARs).

Fig.1.Generalarrangementofa100‐m²EDGGBiodiversity
PlotandthetwoseriesofnestedsubplotsinitsNWandSE
corners.Toestablishthe100m²asaprecisesquare,firstthe
NE‐SWdiagonalof14.14misdelimited(Drawing:I.Dem‐
bicz).
A.3Allplotshaveasquareshape.Somewidespreadmulti‐
scalerecordingschemesusedifferentplotshapesdepending
ongrainsize(e.g.Shmida1984;Stohlgren1995;Peetetal.
1998).However,sinceplotshapesignificantlyinfluencesspe‐
ciesrichness(Stohlgren2007;Bacaroetal.2015;Güleretal.
2016),constantshapeisimportantforcross‐scalestudiesand
analysesofSARs.Amongallthepossibleshapes(squares,
rectangles,circles,hexagons,irregularforms),squaredplots
haveamultitudeofadvantages:(a)apartfromcirclesand
hexagons,theyarethemostcompactform,andthus,onav‐
erage,reflecttheleastpronouncedabioticgradientand
thereforetheclosestlinkbetweenenvironmentalconditions,
speciescompositionandrichness;(b)unlikecirclesandhexa‐
gons,squareplotscaneasilyandpreciselybedelimitedinthe
fieldwithlittleeffortand(c)smallsquarescanbeaggregated
tolargerones,whichisnotpossibleforcirclesorhexagons.
Whilecircles(e.g.Jonssonetal.1992;Olanoetal.1998;
Szwagrzyketal.2001)andhexagons(e.g.Jurasinski&
Beierkuhnlein2006)mightbebeneficialforveryspecificsam‐
plingpurposes,weconsiderthesquaretobethemostpracti‐
calshapeformulti‐purposephytodiversitysamplingap‐
proaches,alsoconsideringthatthegreatmajorityoflegacy
datahasalsobeenrecordedonplotsofthatshape.Inprac‐
tice,the100‐m²plotisestablishedfirstbymeasuringadiago‐
nal(14.14m),markingthetwocornersnottobeusedforthe
nested‐plotseries,fixingafibreglassmeasuringtapeat0m
andat20matthesetwocornersandpullingitatthe10‐m
markuntilbothsidesarestraightlines(Fig.1).Accordingto
ourexperiences,wedonotrecommendmetalmeasuring
tapesastheyaretoostifftoallowprecisedelimitationofthe
squaresinthecorners.Alsothe10‐m²plots(3.16medge
length)arebestdelimitedusingfibreglassmeasuringtapes
andmetalpegs,whilefor1m²(1medgelength)and0.1m²
(0.32medgelength)itismoreconvenienttobenda2‐mfold‐
ingruleatarightangleandtolayitontheground.Forthe
threesmallestgrainsizes,0.01m²(0.1medgelength),0.001
m²(0.032medgelength)and0.0001m²(0.01medge
length),inmanycasesthebestwayisnottolay‐outtheinner
margins,butjustdirectlymeasurethepositionofplantsthat
16 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo5.EDGGBiodiversityPlotinanalpinesteppeof
Mt.Damavand,Iran,2016(Photo:A.Talebi).
Photo6.EDGGBiodiversityPlotduringanadvancedstudent
fieldcourseinNEBrandenburg,Germany,2016.Thestudent
groupinthebackgroundisdetermininggrasshoppersthat
justhavebeencollectedonthediagonalofthe100‐m²plot
(Photo:J.Dengler).

17
arepresumablyclosetothenon‐markedinnermarginsfrom
theoutermarginsthataremarkedwiththemeasuringtape
anyway.

A.4Theplots<100m²arenestedandreplicatedtwicein
twooppositecornersofthe100‐m²plot(Photo6).Since
relativevariabilityofspeciesrichnessandofpracticallyany
otherrelevantparameterincreasestowardssmallerplot
sizes(seeDengler2006),itisimportanttoreplicatethegrain
sizesbelowthelargestones.Foranalysesofspecies‐area
relationships,itisbeneficialtousetheaveragevaluesofthe
replicates,whileusingjustoneplotpergrainsize(e.g.Löbel
2002;Dolnik2003)cansignificantlydistortresults(Dengler&
Boch2008).Whilethestandarderroroftheestimatesfor
grain‐sizerichnessvaluesdecreaseswiththenumberofrepli‐
cates,itturnedoutduringtheEDGGfieldpulsesthatusing
onlytworeplicatesisagoodcompromisebetweenprecision
andtimeefficiency.Practically,thetwosubseriesofnested
plotsareplacedintheNWandSEcornersofthe100m²plot.

A.5Theplotsarenormallyorientedalongthecardinaldirec‐
tions(deviationsarerecorded);GPScoordinatesarere‐
cordedindecimaldegrees(WGS84)fromtheNWandSE
cornerofthe100‐m²plot,usingtheaveragingfunctionto
achievethebest‐possibleprecision(since2009),andthese
cornersarepermanentlymarkedwithburiedmagnets
(introducedafterfieldpulse2016).Thesemeasuresare
aimedatenablingfuturere‐visitationwithprecisere‐location
ofthesameplots.Withthisminimaladditional“investment”
oftimeandmaterial,theEDGGBiodiversityPlotsbecome
realpermanentplots,makingthemthebestpossiblesolution
tostudyvegetationdynamicswithoutanydistortions(i.e.
pseudo‐turnover)throughinaccuratere‐location(seeChytrý
etal.2014).

A.6InadditiontotheEDGGBiodiversityPlots(100m²),
“normalplots”of10m²aresampledwiththesameparame‐
tersasthe10‐m²subplotsoftheBiodiversityPlots(seebe‐
low),butwithnonesting.Theseplotsaremuchlesstime‐
consumingthanEDGGBiodiversityPlotsandtheadditional
samplingof“normalplots”allowshigherreplicationandbet‐
tercoverageofenvironmentalgradientsforthismajorgrain
size.Sincethenormalplotsareineveryrespectidenticalto
the10‐m²subplotsfromthecornersoftheEDGGBiodiversity
Plots,theycanbecombinedinoneanalysis,whichimproves
thestatisticalpoweroftheanalysesat10m²(seeexamples
inTurtureanuetal.2014;Kuzemkoetal.2016;Polyakovaet
al.2016),andensuresthat,despitelimitedsamplingtime,
enoughplotsarerecordedformeaningfulvegetationclassifi‐
cation(Dengleretal.2012a;Pedashenkoetal.2013;
Kuzemkoetal.2014).

B.Speciesrecording

B.1Alllivingterricolous(i.e.soil‐dwelling)vascularplants,
bryophytes,lichensandmacro‐“algae”arerecorded.Be‐
sidesvascularplants,wealsorecordallotherphoto‐
autotrophicterricoloustaxathataremacroscopicallyvisible,
meaningthatweaimatgeneratingacompletepictureofthe
vegetation.Bryophytesandlichenscancontributeverysub‐
stantiallytotheoverall“phytodiversity”(acknowledgingthat
lichenstaxonomicallyarenotplantsbutsymbiosesoffungi
withphotoautotrophicpartners)ofgrasslands(Dengler2005;
Mülleretal.2014;Bochetal.2016;Dengleretal.2016).
Moreover,multi‐taxonstudiesaregenerallyveryinsightful
(e.g.Zulkaetal.2014;Manningetal.2015)andthethree
maintaxonomicgroupsofvegetation:vascularplants,bryo‐
phytesandlichens,showquitecontrastingrelationshipsto
environmentaldrivers(Löbeletal.2006;Lenoiretal.2012;
Polyakovaetal.2016).Inpractice,deadmaterialofperennial
plantsisnotconsidered,whiledeadannualsfromthesame
yeararerecordedwhenpresent.Wedothisbecausewecon‐
siderthatarecordofaplantcommunityshouldreflectacom‐
pleteyear,notjustaseason(Dengler2003).Theoretically,a
bettersolutionwouldbetworecordingsperyearincommuni‐
tieswithapronouncedspring‐ephemeroidaspectandcom‐
biningbothrelevésintoone(Dierschke1994),butthisisim‐
practicalforaone‐timefieldpulse.

B.2Presences‐absencerecordingwiththeshoot‐presence
systemforallplotsizes.Therearetwocommonwaystore‐
cordplantspeciespresenceinplots,rootedpresence(similar
tobutnotidenticalwiththegrid‐pointsystem)andshoot
presence(any‐partsystem)(Williamson2003;Dengler2008).
Whileforlargergrainsizestheresultsofbothmethodsdiffer
onlynegligibly,therichnessrecordedwithrootedpresences
deviatesmoreandmorenegativelyfromshootpresenceval‐
ueswithdecreasingplotsizes,whichistheoreticallyobvious
(Williamson2003),buthasrecentlyalsobeendemonstrated
empiricallyforgrasslands(Güleretal.2016;Cancellierietal.
2017).Therefore,dataderivedfrombothmethodscannotbe
directlycompared.Wedecidedtouseshootpresencebe‐
cause(a)thismethodisadvantageouswhenanalysingSARs
asbothwaysofrecordingnecessarilyshowdeviationsfroma
powerlawatsmallspatialscales,butthesedistortionsare
muchstrongerandoccuratafarlargergrainsizeforrooted
presencethanforshootpresence(Williamson2003;Dengler
2008)and(b)shootpresence,i.e.assuminganindividualas
occupyinganareaandnotonlythepointwhereitpenetrates
thesoilsurface,betterreflectswhichspeciesareinteracting
inthestudiedplot.Inpractice,recordingshootpresenceis
challengingforthethreesmallestgrainsizesof1cm²,10cm²
and100cm².Hereitisimportantthattheobserverisvery
carefulnottodistorttheoriginalarrangementofthevegeta‐
tionwhenplacingthepegsinthecornersandestablishingthe
plots.Forthesesmallestplots,asingleobservershoulddothe
Bulletin of the Eurasian Dry Grassland Group 32 October 2016

recording.Theobservershouldalwayslookfromthesame
angleintotheplotandstartrecordingplantsfromthehigh‐
esttothelowestlayers,withoutrecordingadditionalplants
ofthehigherlayeraftertheyhavebeenbentawaytosample
plantsofthelowerlayers,noraftertheplotshavebeenaf‐
fectedbywind,etc.Ourexperiencesuggeststhatrepresenta‐
tiveresultscanbeachievedwiththeshootpresencemethod
ifoneobserverworksfastandthoroughly.

B.3Additionalpercentagecoverestimationsforthe10‐m²
plots.Traditionally,phytosociologistsrecordedplantper‐
formanceinaplotwiththecombinedcover‐abundancescale
ofBraun‐Blanquet(1964)oroneofitsmanymodified/
refinedversions(e.g.Wilmanns1998).Thisapproachhas
multipleshortcomings,inparticularthecombinationoftwo
differentcriteria,coverandabundancewhich,inthestrict
sense,precludesmostmathematicalanalysesbutwhichis
oftenignored.Furthermore,mostnumericalapproachesdo
notcalculatewiththecover‐abundancescale,butback‐
transformeachcover‐abundanceclasstothemeanofits
range,introducingadouble‐erroroftransformation:first
fromwhatisseeninthefieldtoanabstractcategoryand
thenbacktoarealcovervalue,whichinsomecasescanbe
quitedifferentfromtheoriginalvalue.Imagine,forexample,
acoverof5%,whichbelongstothetraditionalBraun‐
Blanquetcover‐abundancecategory“2”,whichthenusually
isback‐transformedtoacovervalueof15%(because2
standsfor5−25%),meaningthatthisstepintroducedathree
‐folderror.Lastbutnotleast,theBraun‐Blanquetscaleand
almostallsimilarscalesaretoocoarseforrecordingspecies‐
richgrasslandsofhighevenness(wherealmostallspecies
areinthecategory2mor2a)orinverysparsevegetation
(wheremostspecieshavelessthan1%cover.However,itis
abigdifference,i.e.afactorof10,000,whetherthecoveris
0.0001%or1%,whichisnotreflectedintraditionalscales).
Tofacilitaterealisticcoverestimates,we(a)use“estimation
aids”suchasthecalculationtowhichfullyfilledsquaretypi‐
calcoverpercentageswithina10‐m²plotwouldcorrespond
(Table1)and(b)adviseparticipantsthattheyshouldalways
double‐checkthatthecumulativecoverofspeciesofone
groupisatleastashighastheindependentlyestimated
coverofthatgroup.

C.Structuralandenvironmentalvariables(ineach
10‐m²plot)

C.1Coverofvegetationlayers:Coverofthetree(woody>5
m),shrub(woody0.5−5m),herb(woody<0.5mandherba‐
ceous)andcryptogamlayersareestimatedaspercentages
(since2009).Additionally,theherblayerissubdividedinto
thefunctionalgroupsphanerophytes,chamaephytes,grami‐
noids,legumeforbsandotherforbs,allowingforoverlap
betweenthese(adoptedafter2016).Thislaststepdoesnot
onlyprovidevaluabledatainitself,butalsoallowsforcross‐
checkingtheconsistencyofspeciescoverdata(seeB.3).

C.2Maximumheightoftree,shrubandherblayers.

C.3Measurementof“standardheight”ofthevegetation
(since2016;prototypeduringfieldpulse,improvedversion
afterwards;Photo7):Atfiverandompointsintheplot,a
circularplasticdiscwithacentralborehole(22.5cmdiame‐
ter,117g)isreleasedalongtheinvertedpenetrometer(see
below),thehandleofwhichisplacedontheground.The
heightwherethefallingdiscisstoppedbythevegetationis
measuredattheborehole.Thefivemeasurementsprovidea
reproduciblemeasureoftheheightatwhichthevegetation
becomesdense,aswellasofitsspatialvariability.
18 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo7.Usingapreliminaryversionofthediscforstan‐
dardisedassessmentofvegetationheightduringthe
EDGGFieldWorkshopinSerbia2016(Photo:I.Dembicz).
Table1.Areasofcompletelyfilledsquaresthatcorrespond
tocertainpercentagecovervaluesin10‐m²plots.
Percentage
cover value
Area in
m2
Area in
cm2
Edge length of
square in cm
5 0.5 5000 70.7
4 0.4 4000 63.2
3 0.3 3000 54.8
2 0.2 2000 44.7
1 0.1 1000 31.6
0.5 0.05 500 22.4
0.1 0.01 100 10.0
0.05 0.005 50 7.1
0.01 0.001 10 3.2
0.005 0.0005 5 2.2
0.001 0.0001 1 1.0


C.4Abovegroundbiomass(firstvariantsduringfieldpulse
2015;currentversionafterthefieldpulse2016;Photo8):
Withineach10‐m²plot,wecliptheabovegroundbiomass
withintworandomareasof20cm×20cmtothesoilsurface.
Wethenpoolbothsamples,i.e.atotalsurfaceof800cm²,
afterwhichwedryandweighthem.Samplingtwoseparate
areasallowsamuchbetterestimateofthemeanbiomassper
1m²withinthe10m²plotthanasingleplotwoulddo(asin
2015).Duetotherelativelysmalltotalsurface,theamountof
biomassisstillpracticable,evenduringlongerfieldpulses,i.e.
thematerialcanbetransportedandpre‐dried.Whilein2015,
weseparatedthethreefractionsintolivingvascularplants,
livingnon‐vascularplantsandlitter,since2016wehavetaken
justonecombinedsampleforbiomasss.l.becausetheprevi‐
ousapproachwastootime‐consuming.Practically,thearea
tobesampledcanbedelimitedbyaspecificallymanufactured
steelframeormoreeasilybyaframecreatedbybendinga
foldingrulerfourtimes.Theinneredgeis19cm,butsinceitis
impossibletofixtheposition100%duringbiomasscutting,
thisisagoodapproximationoftheintendedsize.Dryingis
doneinanovenat65°Cuntiltheweightremainsconstant.

C.5Coveroflitteranddeadwood:Percentagecoverafter
virtuallyremovingallvegetation.Notethatthewidespread
approachofphytosociologiststoestimateonlythatpartof
thelitterthatisvisiblefromabove,i.e.notcoveredbyliving
vegetation,precludesusinglittercoverasapredictorofvege‐
tationattributes,becausethetwovariableswouldthennot
beindependentofeachother.

C.6Fractionsofabioticsoilsurface:Percentagecoverofthe
threetextureclasses:stonesandrocks(diameter>63mm),
gravel(2−63mm)andfinesoil(<2mm)atthesoilsurface
aftervirtuallyremovingallvegetation,litteranddeadwood,
thus,summingupto100%cover.Notethatthewidespread
approachofphytosociologiststoestimateonlythatpartof
thesoilsurfacethatisvisiblefromabove,i.e.notcoveredby
livingordeadvegetation,precludesusingthesefractionsas
predictorsofvegetationattributesbecausetheywouldthen
notbeindependent.

C.7Slopeaspectandinclination:Practicallymeasuredby
placingthepenetrometer(seebelow)onthegroundalong
theslopeline.Aspectismeasuredindegreeswithacompass
and(mean)inclinationindegreeswithaninclinometer.
Nowadays,smartphoneappsareavailablethatdobothina
veryconvenientwaywhenplacingthesmartphoneonthe85
cmlongpenetrometer.

C.8Microrelief:Isdefinedasthemaximumdistancetothe
groundwhenplacingthepenetrometer(seebelow)tothe
groundinthemostruggedpartoftheplot,measuredper‐
pendiculartothedevice(since2014;Fig.2).Formerly,we
tookthismeasurementplumb‐vertical,butthisapproach
stronglyconfoundedmeasurementsofmicroreliefbyslope
inclination.

C.9Soildepth:Ismeasuredatfiverandompoints(toallow
calculationofmeanandstandarddeviation)usingoursoil
depthindicator(penetrometer;Photo9).Thisisasteelpole
of85cmlengthand1.0cmdiameter,pointedatoneendand
withahandleattheother.Itispushedintothegrounduntilit
hitsarockorthesoilbecomessodensethatitcannotbe
pushedfurther.Eachdepthmeasurementisnotedsepa‐
rately,evenifitis“0cm”(rockatthesurface)or“>80
cm”(noresistanceatanydepth).Itisobviousthatthismeas‐
urementshouldpreferablyalwaysbedonebythesameper‐
sonofaverageweightandstrength.The“odd”lengthofthe
penetrometerisbecausethiswasthelengthofourfirstde‐
vice.However,itturnedoutthatadeviceofthislengthstill
canbereasonablywellcarriedinchecked‐inluggageduring
airtravel,whilealengthof1mwouldalreadycauseprob‐
lems.

19 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo8.ClippingbiomassduringtheEDGGFieldWorkshop
inSerbia2016(Photo:J.Dengler).
Fig.2.Illustrationofhowtomeasurethemaximummicrore‐
lief(orangeline)inaplot(inourcasethe10m²plots)
(Drawing:I.Dembicz).

C.10Soilsamples:Amixedsoilsampleoftheuppermost10
cmofthemineralsoilistakenfromfiverandomlocations
withinthe10‐m²plot(Photo10).Thissampleisair‐dried
duringthefieldpulse(Photo11)andafterwardsdriedat65°
C.Fromthissample,wedetermineasaminimumthefollow‐
ingparameters:(a)skeletoncontent(i.e.massfractionof
particles>2mm),(b)textureclass(mostlyestimatedwitha
fingertest−seeSchlichngetal.1995;Ad‐hoc‐AGBoden
2005–sometimesmeasured,whichistime‐consumingand
costly);(c)pH(inasuspensionof10gdrysoilin25gaqua
dest.);(d1)humuscontent(aslossatignitionat430°Cuntil
constancy)or,ifresourcesallow,(d2)CandNcontents
(withaC/Nanalyser),includingcorrectionforCfromcar‐
bonates.
C.11Land‐use:Isproblematictoassessduringaone‐off
visit.Wetrytocategorizeeachplotbasedontraces,suchas
faeces,grazingmarks,presence/absenceofpastureweeds,
intopasture(i.e.livestockgrazed),meadow(i.e.mown)or
un‐usedinrecentyears(abandonedsemi‐naturalgrassland
ornaturalgrassland)(e.g.Turtureanuetal.2014).Addition‐
ally,weuseburningtracestodecidewhethertheplotwas
burnedduringthecurrentyearornot.Anymoreprecise
informationonland‐use,themanagementregimes,their
timingandduration(e.g.livestocktype,numberofanimals,
combinationofmowing,grazingandfertilization,peculiari‐
tiesingrasslandhistory,etc.)thatisavailableisrecorded.
Unfortunately,ourexperienceisthatduringaone‐offvisit
suchdatacanhardlybegatheredconsistently,sothatin
noneofthefieldpulsessofarwereweabletousemore
detailedland‐useparametersforanalyses.

D.Datamanagement

Tofacilitateandstandardisedatacollectionandmanage‐
ment,theEDGGprovidesandregularlyupdatesaseriesof
documents,i.e.instructions,templatesforprintedforms
andspreadsheets,thecurrentlyup‐to‐dateversionsof
whichaccompanythisarticle.Allthesedocumentsareopen
accessandcanbemodifiedaccordingtopersonalneeds.
OnlineResource1containsadetailedlistofequipment
neededforsamplinglikethatdoneduringEDGGfieldpulses,
dependingonthedurationandnumberofparticipants
(Photo12).OnlineResource2providesdetailedpractical
instructionsonhowtoimplementtheEDGGsamplinginthe
field,whileOnlineResource3describesthedatahandling
andrecordingafterthefieldwork.OnlineResources4and5
arethecurrenttemplatesforbiodiversityplotsandfornor‐
malplots.OnlineResources6and7,finally,arespread‐
sheets(*.xlsxformat)fortheefficientdataentryofspecies
dataandheaderdata,respectively.Theyincludesomeem‐
beddedfunctionsthatfacilitateworkandprovidesomesim‐
pledatachecks(fillinginthespecieslistfor100‐m²plots
automaticallybasedonthetwocorners,checksofconsis‐
tencyofcovervalues,calculationofmeanandstandardde‐
viationforparameterswithmultiplemeasurements,de‐
scriptivestatisticsforparametersacrossallplotstocheckfor
outliers/entryerrors).Fromthesetwospreadsheets,the
relevantdatasetsforthemultipleanalyses,beitinRorany
otherstatisticalsoftware,canbederivedwithafewclicks.
ThedataoftheEDGGfieldpulsesandsomerelatedsam‐
plingschemesarestoredinacommondatabase,registered
intheGlobalIndexofVegetation‐PlotDatabases(GIVD;
Dengleretal.2011)asEU‐00‐003(Dengleretal.2012b).
Thesedataareavailableforcommondataanalysesbythe
contributorsandtheirpartners.Moreover,thefieldpulse
dataofthe10‐m²plotsarecontributedtoexistingnational
orregionalpartnerdatabasesoftheEuropeanVegetation
Archive(EVA;Chytrýetal.2016)andtheglobalcounterpart
“sPlot”(Purschkeetal.2015)sothattheyareavailablefor
continentalorglobalanalyses.

E.Possibleextensionsofthemethodology

E.1Otherspatialscales:Themostmeaningfuladditional
scalewouldbe1000m²(31.62m×31.62m)sincethisisa
commongrainsizeinmanybiodiversitysamplingschemes
worldwide,albeitmostlyrealisedas50m×20m(e.g.
20 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo9.Applyingthepenetrometertodeterminetheac‐
cessiblesoildepthduringtheEDGGFieldWorkshopinNa‐
varre,Spain,2014(Photo:J.Dengler).

Shmida1984;Peetetal.1998;Jürgensetal.2012).Thereare
severalwaystoarrangenestedplotswithin1000m²inaway
thatiscompatiblewiththeEDGGBiodiversityPlots:(a)place
oneEDGGBiodiversityPlotinthecentreofthe1000m²plot;
(b)placetwoEDGGBiodiversityPlotsintwooppositecor‐
ners;(c)placesinglenestedseriesof0.0001−100m²intwo
cornersor(d)thevariantshowninDengler(2009:Fig.1).
Oneshouldbeawarethatadding1000m²drasticallyin‐
creasesthetimeneededforsampling(comparethetimesfor
asinglenested‐plotseriesupto900m²asreportedbyDolnik
2003).Therefore,oneshouldonlyoptforthisadditionwhen
thereareadequateresourcesavailabletosamplethe1000
m²ascomprehensivelyasthe100m².Mostconvenientlythis
canbedoneinrelativelyspeciespoorvegetationwithlow
covervalues,e.g.insomeopenherbaceousvegetationof
SouthernAfrica(Jürgensetal.2012)orintransitionsfrom
steppestosemi‐desertsinIran(whereagroupincludingA.N.
iscurrentlydoingthis).Samplingsmallergrainsizesthan1
cm²,i.e.1mm²and10mm²,isalsopossible,butrequiresa
specialdevice(Dengleretal.2004).Finally,itcanmakesense
to“insert”additionalplotsizeswithfullsampling(including
covervaluesandenvironmentaldata),suchas16m²or25
m²,ifthisisanationalstandardforsamplingherbaceous
vegetationforphytosociologicalpurposes.Inthiscase,how‐
ever,thisadditionalplotsizeshouldnotbeusedinSAR
analyses,toavoidbias.

E.2Higherreplicationatsmallerscales:Sincethecoefficient
ofvariationofspeciesrichnessstronglyincreaseswithde‐
creasingplotsize(Dengler2008andreferencestherein),
increasinglymorereplicateswouldbenecessaryforsmaller
plotstoestimatemeanspeciesrichnesswiththesamepreci‐
sion.Thus,theoriginalapproachofDengler(2009)proposed
thattowardseachsmallerscalewithinthe100‐m2plots,and
downto0.01m2,thenumberofsub‐plotreplicatesisdou‐
bled.Duetotimeconstraintsandbecauseitishardlypossi‐
bletoarrangesuchanincreasingreplicationthatisboth
nestedandunbiasedwithrespecttothe100‐m²area(i.e.
doesnothavehighersamplingintensityinsomeregionsthan
inothers),thisapproachwasneveradoptedduringtheEDGG
fieldpulses.However,Dengleretal.(2004)andBoch(2005;
seeDengler&Boch2008)usedfourandfiveseriesofnested
plots(0.0001−10m²)withinthe100‐m²plot.RecentlyCan‐
cellierietal.(2017)adoptedtheideaofDengler(2009:Fig.
2),althoughwithalimitednestedseriescomposedofonly
threespatialscales.

E.3Stratified‐randomsampling:StepA.1oftheEDGGsam‐
plingmethodologyisaimedatapproximatingadatasetsimi‐
lartoonegainedwithstratifiedrandomsampling,butwhen
suchanaprioristratificationisnotfeasibleduetotimecon‐
straintsorlackofsuitableinformationlayersforuseina
GeographicInformationSystem(GIS).Basically,thesampling
approachofDengler(2009)isapplicableinsubjectivelyde‐
limitedhabitattypes,withstratifiedrandomsampling(oran
21 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo10.TakingasoilsampleduringtheEDGGField
WorkshopinKhakassia,Russia,2013(Photo:J.Dengler).
Photo11.Dryingandsortingofsoilsamplesduringthe
EDGGFieldWorkshopinNavarre,Spain,2014(Photo:J.
Dengler).
Photo12.EquipmentneededfortheEDGGsamplingmethod‐
ology,spreadtobepackedbythedifferentteamsduringthe
EDGGFieldWorkshopinSerbia,2016(Photo:C.Marcenò).

approximationofit)orwithfullyrandomsampling(forthe
prosandconsofthesesamplingapproaches,seeWildi1986).
OnlyfuIlyrandomsamplingallowscalculationoftruespatial
meansofattributes,suchasspeciesrichness(e.g.Dengler&
Allers2006;seealsothegrid‐basedrandomapproachbyCan‐
cellierietal.2017,whichwasinspiredbyChiaruccietal.
2012),butthisusuallyleadstoastrongunderrepresentation
ofrarehabitattypes(Diekmannetal.2007).Stratified‐
randomsamplingtheoreticallyallowsonetogetadataset
thatismorebalancedwithregardtoenvironmentalgradients
(thanfullyrandomsamplingwould)andeventoapproximate
spatialmeans(whentakingthefractionalextentofthestrata
intoaccount),whileavoidingthepotentialbiasesofsubjec‐
tivelylocatingtheplots.However,stratified‐randomsampling
requiresthatthemainenvironmentalgradientsarerather
clearaprioriandavailableasGISlayersforthestudyregion,
whichisnotusuallythecaseforEDGGfieldpulses,oneof
whosemainaimsistostudyundersampledregions.Ifthe
prerequisitesaremet,werecommendconsideringastratified
‐randomapproach(andaimtoimplementitintheField
Workshop2017inCentralItaly;seeFilibecketal.2016).This
approachmeansthatrandomcoordinateswithineachlevel
ofoneorseveralcrossedmainenvironmentalfactorsare
generatedwithinaGISandthensampledinthefield.Forex‐
ample,onecouldstratifytheregionbyelevationandbedrock
typeorbyland‐usetypeandslopeposition.Itisself‐evident
thatoneneedstodecideforone,twooramaximumofthree
gradients,eachsubdividedintoasmallnumberofcategories,
becauseotherwisethenumberofplotsnecessarywouldsoon
becomeunrealistic.Oneshouldhoweverbeawareofthe
potentialproblemsofastratified‐randomapproach,even
whentheseprerequisitesaremet.Ontheonehand,thea
prioriassumptionaboutthemaingradient(s)mightturnout
tobewrongandthenthesamplingwouldnotbeoptimal.For
example,Baumannetal.(2016)usedEDGGBiodiversityPlots
withelevationalstratification,onlytofindoutthattheeleva‐
tionalgradientintheircasewasofsubordinateimportance
forthespeciesrichnesspatterns.Ontheotherhand,strati‐
fied‐randomsamplingsignificantlyincreasesthetimeneeded
tofindandreachtheplotsinthefield,whichinsomecases
mighteventurnouttobeimpossibleduetoinaccessibility.

E.4Betterassessmentofbetadiversity:Throughthemulti‐
scalesampling,theEDGGBiodiversityPlotsprovidea
straightforwardtooltoassessbetadiversityatthesmallest
scales(i.e.within100m²).AsPolyakovaetal.(2016)demon‐
strate,thez‐valuesofthepower‐lawSARsareameasureof
standardised,multiplicativebetadiversity,whichallowscom‐
parisonsofwithin‐plotspeciesturnoverbetweenEDGGBiodi‐
versityPlotsofdifferentecologicalconditionsorregions(see
alsoDengler&Boch2008;Turtureanuetal.2014).Assessing
betadiversityacrosslargerspatialextentsthan100m²,for
exampleacross1km²or1000km²inacomparablemanner,
isnotstraightforwardwiththeEDGGsamplingmethodology
becausebetadiversityvaluesarelargelydeterminedbythe
spatial(andecological)extentofthestudyuniverse(Chiarucci
etal.2009).Theapproachofconstrained‐rarefactionoffersa
waytomakedatafromstudyregionsofdifferentspatialex‐
tentcomparable(Chiaruccietal.2009).However,eventhis
wouldnotaccountforpotentiallydifferentecological/
syntaxonomicaldelimitationsofthe“studyuniverse”indif‐
ferentfieldpulses(if,forexample,inoneonlyFestuco‐
Brometeaweresampledandinanotheralltypesofsemi‐
naturalgrasslands).Therefore,iftheassessmentoflandscape
‐scalebetadiversityisamajoraimofastudy,oneshouldcon‐
sidertheappropriateplacementoftheEDGGBiodiversity
Plots.Oneshoulddecideonthelandscapescaleofinterest,
e.g.200km²(acirclewitharadiusof7.98km),inwhichthe
plotsshouldbelocatedrandomly.InthisrespectE.4canwell
becombinedwithE.3.Ifalessformal,adhocsolutionisre‐
quired,onecouldthinkofplacingasetoffive(oranother
fixednumber)EDGGBiodiversityPlotswithinthesurveype‐
rimeterhaphazardly,withtheonlyrestrictionthateachof
theseshouldrepresentadifferentgrasslandtypeor,ifonly
onetypeisconsidered,comefromadifferentgrassland
patch.

E.5Non‐terricoloustaxaofthevegetation:Alsosaxicolous
(speciesgrowingonrocks),lignicolous(speciesgrowingon
deadwood)andepiphytictaxa(speciesgrowingonthebark
orevergreenleavesofotherplants)belongtotheoverall
phytodiversity.Therefore,werecommendtosamplealso
thesetaxa.Particularly,saxicolousbryophytesandlichenscan
contributesignificantlytotheoverallrichnessinrockygrass‐
lands(e.g.Boch2005;Bochetal.2016).Unfortunately,such
samplingrequiresspecificequipment(e.g.aknifetocollect
lignicolousandepiphyticspeciesaswellasahammeranda
chiseltocollectsamplesofsaxicolouslichensthatcannotbe
identifiedinthefield)andspecialexpertiseintheidentifica‐
tionofthesespecies.Fornon‐experiencedobserversaiming
atsamplingnon‐vascularplants,onepossibilitymightbethe
samplingofsocalledmacro‐cryptogams,whichareeasily
discernibleinthefield(e.g.excludingcrustoselichensand
verysmallbryophytes).Theirrichnesscanbeusedasanindi‐
catorfortheoverallrichnessofcryptogams(Bergaminietal.
2005).

E.6Animaltaxa:Giventhehighpotentialvalueofmulti‐
taxonstudiestounderstandpatternsanddriversofbiodiver‐
sity(e.g.Allanetal.2014;Zulkaetal.2014;Manningetal.
2015;Soliveresetal.2016),itishighlydesirabletoalsosam‐
pleanimaltaxa,forwhichsamplingatthegivenspatialscales
makessenseandcanbeperformedduringasinglevisit.In
anycase,itmustbedecidedwhichofthegrainsizescanbe
sampledmeaningfully,giventhatanimals,unlikeplants,are
ontheonehandmobileandontheotherhandnotalways
discernibleeveniftheyarepresent(i.e.recordsusuallyrepre‐
sentactivity,notpresence).Incontrasttoplants,typically
onlyoneorperhapstwoofthestandardgrainsizescanbe
sampledandmatchedwiththephytodiversitydata.Vegeta‐
tionstructureand,insomecases,plantspeciescomposition
andrichness,haveastronginfluenceonspeciescomposition,
richness,activitiesandabundancesofmanyaboveand
belowgroundinvertebratetaxa(Lawton1983;Borges&
Brown2001;Birkhoferetal.2011;Simonsetal.2014).For
example,spidersaspredatorsareinfluencedindirectlyby
22 Bulletin of the Eurasian Dry Grassland Group 32 October 2016

changesinmicroclimaticconditions,preyabundance,sites
forbuildingwebs,shelteringand/oroviposition(Gunnarsson
1990;Halajetal.1998;McNett&Rypstra2000).Therefore,
collaborationofbotanistsandzoologistsinbiodiversityas‐
sessmentsishighlydesirable.
DuringtheFieldWorkshopinNavarre,Spain,vegetation‐
dwellingspiders(Araneae)weresampledbyN.Y.P.onthe
100‐m²plotswithstandardsamplingmethods,suchassweep
‐nettingandhandcollecting(Duffey1974;Photo13).Asthe
biodiversityplotsarerelativelysmallforsweep‐netting,one
sampleof15sweepswastakeninsideagivenplotandthree
samplesadjacenttoit.Usingrepeatedsweepsatthesame
plotisineffectiveforspiders,astheyfallonthegroundand
donotascendintothevegetationagainimmediately.Some
biodiversityplotscouldnotbesampledbecauseofanex‐
tremelylowsward.Spiderdiversitydatahavenotbeenana‐
lysedyet,butthissamplinghasledtothedescriptionofa
newspiderspecies(Kastryginaetal.2016).Whereitispossi‐
ble(butthisisevidentlynotthecaseduringEDGGfieldpulses
withonlyonevisitpersite),werecommendtostartinearly
springwithpitfalltrapping,usingatleastfivetrapspereach
EDGGBiodiversityPlot.Thetrapexpositionmayvaryfrom5–
10daysmonthlytoamonth.Spidersarerecommendedtobe
collectedduringthewholevegetationperiodbecauseofthe
differentseasonalactivitiesandmaturationtimes.Ifthereis
noopportunitytoconductsuchlong‐termstudies,itispossi‐
bletolimitthemtospringandearlysummer.Sweep‐netting
iseffectivefromlatespringuntilmid‐summer.Thesuction
methodusingaTullgrenfunneliseffectiveforplotswithlow
vegetationinthesameperiod.Ingeneral,itispreferableto
addoneautumnalsample.Quadratsamples(hand‐collecting
in25cm×25cmplotsonthegroundand/orlitterlayer)can
reveallessmobileground‐dwellingspecies.Thismethodcan
beparticularlywellcombinedwiththevegetationdata,as
onecantakeasmanysamplesasrequiredinsideavegetation
plot.Toestimatethecompletespidercommunityinagiven
EDGGBiodiversityPlot,onewouldneedtocombineallthe
above‐describedmethods.
Recently,studentsofthefirstauthor,includingB.H.,sampled
grasshoppers(Orthoptera)inEDGGBiodiversityPlots(Photo
14).Giventherelativelylownumberofspecies,Orthoptera
areeasiertoidentifythanmanyotherinsecttaxa.Phyto‐
phagousOrthopteraareusuallypolyphagous,meaningthey
arenotlimitedtojustonefamilyoffoodplantsbutrather
consumeawiderangeofplantspeciesacrossdifferentplant
families.Therefore,itcanbeassumedthattheoccurrenceof
certainspeciesofOrthopterawillnotdependontheoccur‐
renceofcertainplantspecies,butwillratherfollowfactors
likethemicroclimate,vegetationstructure(Gardineretal.
2002),plant‐coverorland‐use.ThesamplingofOrthoptera
shouldtakeplaceduringwarmsunnydaysinlatesummer
(August−September)toensuredeteconofmostlyimagines
(whichareeasiertodeterminethanjuveniles).Daysofsam‐
plingshouldnotfollowadayofintenserainfall.Weusedthe
sweep‐nettingmethodbecauseitisthemostrapidmethodin
thefieldanddoesnotrequireexpensiveequipment.The
mostcommonlyusednetsizeis38cmdiameter(Bomar
2001;Gardineretal.2005).Withineach100‐m2EDGGBiodi‐
versityPlot,firsttheNE‐SWdiagonal(i.e.theonethroughthe
cornerswithout10‐m²vegetationplots)wassweptbyad‐
vancingonestepforwardaftereachsweep.Bydoingso,the
sweepingofthediagonalwascompletedafterabout15con‐
secutivesweeps.Inaddition,thewhole100‐m2plotwassam‐
pledagainthreetimesbywalkingaroundandsweeping
withintheplotforaboutfivesecondseachtimetoensure
thatthewholeplotareawassampled.Aftereachsweepthe
Orthopteracaughtinthenetweretransferredintoplastic
boxesforsubsequentidentificationandcounting.Whilethe
describedmethodgenerallyworkedwell,itbecomesprob‐
lematicinvegetationplotswithtallervegetation(>50cm
plantheight),asthecatchingefficiencymaybeimpededby
vegetationstructure(Gardineretal.2005).
IntheSwissBiodiversityMonitoring,apartfromvascular
plantsandbryophytes,alsolandsnails(Gastropoda)are
sampledonthesame10m²plots(KoordinationsstelleBDM
2014).Otherinvertebrategroupsthatarepotentiallysuitable
Photo14.Samplinggrasshoppersona100‐m²EDGGBiodi‐
versityPlotduringanadvancedstudentfieldcourseinNE
Brandenburg,Germany,2016(Photo:J.Dengler).
23 Bulletin of the Eurasian Dry Grassland Group 32 October 2016
Photo13.Samplingvegetation‐dwellingspidersona100‐m²
EDGGBiodiversityPlotduringtheEDGGFieldWorkshopin
Navarre,Spain,2014(Photo:J.Dengler).

forinclusionintheEDGGBiodiversityPlotsincludeleafhop‐
pers(Auchenorrhyncha)(e.g.Primietal.inpress).

E.7Moreandbetterstandardisedenvironmentaldata:
Clearly,thegreatertheamountofstandardisedabioticdata
thatareassociatedwiththerecordedbiodiversitydata,the
moreanalyticalopportunitiestheyoffer.TheEDGGsampling
methodologyrequiresparametersandmeasurementmeth‐
odsthatgeneratereliabledataduringasinglevisitusinglim‐
itedtimeandresources.Amongsoilparameters,goodcandi‐
datesthatwerecollectedduringsomefieldpulsesbutnot
fixedasstandardyetareelectricalconductivity(EC),whichis
particularlyrelevantwhensamplinginaridareasorsaline
habitats,andH‐ andS‐value,fromwhichcationexchange
capacity(CEC)andbasesaturation(BS)canbederived.Ifthe
EDGGBiodiversityPlotsaredistributedwithinarelatively
narrowregionandrevisitingallofthemwithinoneorafew
daysofconstantdryweatherisfeasible,alsosoilwater‐
content(volumetricorgravimetric)wouldbeavaluablepa‐
rameter.

E.8Qualityassessment:Therearerelativelymanystudies
(seereviewbyMorrison2016)thatmeasuredtheimpactof
observer‐relateddiscrepanciesinvegetationsamplingand
warnagainsttheresultingbiasesinspeciesrichness,cover
estimates,andvisualestimatesofothervegetationfeatures.
However,thisissueisstillsurprisinglydisregardedorover‐
lookedinthevastmajorityofpublishedresearchesbasedon
analysisofplot‐baseddataacrossspatialandenvironmental
gradients.Nevertheless,moststudiesonobserver‐related
errorfoundmeanvaluesofpseudo‐turnover(i.e.ofthedif‐
ferenceinspeciescompositionbetweentwoobservers,or
teamsofobservers,surveyingthesameplot)rangingfrom
10%to30%(Morrison2016).Inastudywithfine‐scaleplots
intemperateEuropeangrasslands,Klimeš