Monitoring change in the abundance and distribution of insects using butterflies and other indicator

19 Pages·2005·0.9 MB·English

Phil.Trans.R.Soc.B(2005)360,339–357
doi:10.1098/rstb.2004.1585
Publishedonline28February2005
Monitoring change in the abundance and
distribution of insects using butterflies
and other indicator groups
J. A. Thomas*
NERCCentre for Ecology&Hydrology, Dorset Laboratory, Winfrith TechnologyCentre,
Dorchester DT2 8ZD,UK
Conservative estimates suggest that 50–90% of the existing insect species on Earth have still to be
discovered, yet the named insects alone comprise more than half of all known species of organism.
With such poor baseline knowledge, monitoring change in insect diversity poses a formidable
challengetoscientistsandmostattemptstogeneralizeinvolvelargeextrapolationsfromafewwell-
studied taxa. Butterflies are often the only group for which accurate measures of change can be
obtained.Fourschemes,usedsuccessfullytoassesschangeinBritishbutterflies,thatareincreasingly
being applied across the world are described: Red Data Books (RDB) list the best judgements of
experts of the conservation status of species in their field of expertise; mapping schemes plot the
changingdistributionsofspeciesatscalesof1–100km2;transectmonitoringschemesgeneratetime
series of changes in abundance in sample populations of species on fixed sites across the UK; and
occasional surveys measure the number, boundaries and size of all populations of a (usually RDB)
speciesatintervalsof10–30years.Allschemesdescribeconsistentpatternsofchange,butiftheyare
to be more generally useful, it is important to understand how well butterflies are representative of
other taxa. Comparisons with similarly measured changes in native bird and plant species suggest
that butterflies have declined more rapidly that these other groups in Britain; it should soon be
possibletotestwhether thispatternexistselsewhere.Itisalsodemonstratedthatextinctionratesin
Britishbutterfliesaresimilartothoseinarangeofotherinsectgroupsover100yearsoncerecording
biasisaccountedfor,althoughprobablylower thaninaquaticorparasitictaxa.Itisconcludedthat
butterflies represent adequate indicators of change for many terrestrial insect groups, but
recommended that similar schemes be extended to other popular groups, especially dragonflies,
bumblebees, hoverflies and ants. Given institutional backing, similar projects could be employed
internationally and standardized. Finally, a range of schemes designed to monitor change in
communitiesofaquaticmacro-invertebratesisdescribed.Althoughdesignedtouseinvertebratesasa
bio-indicator of water quality for human use, these programmes could be extended to monitor the
2010 biodiversity targets of the World Summit on Sustainable Development.
Keywords:biodiversity;extinction;habitatloss;climatechange;2010WorldSummitonSustainable
Development
1. INTRODUCTION decade,particularlyindevelopedcountries,owingtoa
Insects pose an obstacle to the development of a series of recording and monitoring schemes which,
scientifically rigorous process for reporting against the although mainly developed 20–30 years ago (Thomas
World Summit on Sustainable Development’s 2010 1984;Masteretal.2000;Wrightetal.2000),haveonly
targettostrivefor‘asignificantreductioninthecurrent recently yielded decisive results. A parallel advance in
rateoflossofbiologicaldiversity’.AsatanearlierRoyal empiricalknowledgeabouttheprocessesdrivingchange
SocietyDiscussionMeeting,EstimatingExtinctionRates ininsectpopulationdynamics(e.g.Dempster&Pollard
(May et al. 1995), entomological knowledge lags far 1981; Dempster & McLean 1998; Godfray & Mu˝ller
behind that achieved for assessments of change in 1998; Hanski 1998, 1999; Hassell 1998a; Murdoch
vertebratesandplants,orformajorecosystemssuchas
et al. 1998; Thomas et al. 1999a,b, 2001a,b; Royet al.
coral reefs and rain forests (Bibby 1994; Coope 1994;
2001;Warrenetal.2001;Bourn&Thomas2002)plus
Ehrlich 1994; Mace 1994; Thomas & Morris 1994,
assessmentsoftheextenttowhichwell-studiedgroups
1995;Mayetal.1995;Pimmetal.1994).Nevertheless,
ofinsectarerepresentativeofobscureones(Thomas&
information about the scale and patterns of insect
Morris1995;Hambler&Speight1996,2004;McKin-
decline has advanced substantially during the past
ney1999;Thomas&Clarke2004),complementmany
theories (May 1974; Godfray & Hassell 1987,1992;
*Authorforcorrespondence:([email protected]). Gaston1991;Lawton1995;Hassell1998a,b)andtoday
represent an additional, indirect method of measuring
One contribution of 19 to a Discussion Meeting Issue ‘Beyond
extinctionrates:monitoringwildnatureforthe2010target’. change.
339 q2005TheRoyalSociety
340 J.A.Thomas Monitoringchange and distribution of insects
Here I seek to describe and evaluate the main tools Four complementary approaches (Red Data Books
currentlyavailabletoentomologiststoassesschangein (RDBs), atlases, time-series and absolute population
insects and other terrestrial and freshwater censuses) have been developed to assess change in
invertebrates; I also analyse the extent to which butterflies,mostoriginatingintheUK.Althoughsome
change in butterflies (which are often the most—or havebeenappliedtootherinsecttaxa,theusefulnessof
only—practicalinsectgrouptostudyacrosstheworld) each has been tested most thoroughly on butterflies,
arerepresentativeofotherterrestrialinvertebrates,and especially the British fauna, on which the next four
suggesthowcurrentmonitoringandmappingschemes sections focus. In addition, a separate approach of
might be improved to meet 2010 targets. sampling freshwater invertebrates has been developed
Insects are difficult to study because they represent acrossEurope(Wright1994;Alba-Tercedor&Pujante
themostspecies-rich,yetoneoftheleastknown,ofall 2000; Johnson & Goedkoop 2000; Verdonschot &
taxaoflivingorganisms,aproblemthatiscompounded Nijboer 2000), Australia (Davies 2000; Simpson &
by a dearth of skilled entomologists. Although the Norris2000), Canada(Rosenbergetal.2000)andthe
number of describedinsect speciesis uncertain dueto USA (Reynoldson et al. 2000). It employs community
synonymsandthelackofagloballist,mostauthorities richness, usually at the family level, as a predictive
recognize 900000–1000000 named morpho-species, technique for evaluating the ecological status of
representing 56% of all species known on Earth freshwater rivers and lakes. Although primarily
(Groombridge 1992; Anon. 2003). Sensible estimates designed to assess annual variation in water quality
ofthenumberofinsectsyettobediscoveredrangefrom for human use (Wright et al. 2000 and papers
another 1 million to 30 million species (Erwin 1982, therein), it provides an infrastructure that could be
1991),althoughmostpredictaround2–8millionmore extended to monitor changes in biodiversity per se
species (May 1990; Gaston 1991; Stork 1997; (Wright et al. 1993).
Ødegaard 2000). Thus, insects lie in the zone of
maximumspeciesrichnesswhenbiodiversityisrelated
to body size (May 1988; Gaston & Lawton 1988;
Finlay 2004): larger species are fewer due to their 2. INSECT RED DATA BOOKS
comparatively wide niches, the greater resources they Compiling a RDB is usually the first, and often only,
needandtheactivemobilityofmanyofthem,whereas assessment of change in insects for most regions or
free-livingorganismssmallerthan1mm3experiencesuch groups.However,itisimportanttorealizethatthemany
rapid passive transportation and mixing by wind and RDBs for invertebrates (e.g. IUCN 1983; Shirt 1987)
currentsthatspeciationthroughgeographical(asopposed contain major uncertainties compared with those for
toniche) isolation is prevented(Finlay2002,2004). vertebrates and plants and cannot reliably be used
Apart from the few trans-continental migrants, the for the sophisticated analyses of change employed for
majorityofstudiedinsectspeciesformsmall,localized reptiles,birdsandmammals(e.g.Mace1995)without
populations, typically occupying narrow specialized majorcorrectionforbiasandrecordereffort(McKinney
nichesinpatcheslessthan2hainsizeandsosedentary 1999;Thomas&Clarke2004).Forexample,theinsect
thatdistancesof1m–1kmof‘non-habitat’mayposea fauna of the UK are better studied than any other
majorbarriertodispersal(Thomas1991;Thomasetal. comparableareaintheworld(Bratton1991;Collins&
2005).Itfollowsthatmanymeta-populationsofinsects Thomas 1991; Ehrlich 1994) and the British RDB-
aresogeographicallyisolatedwithineachspecies’range insectswasexceptionallythorough,beingbasedonthe
thatregionaladaptationsto,andcoevolutionwith,their recommendations of 95 experts who, over a 5 year
local communities is commonplace, resulting in period, reviewed their specialist insect groups (apart
regional sets of functionally distinct subspecific forms from Microlepidoptera, Hymenoptera Parasitica and
(Thompson 1994), including, many believe, local someDiptera)usingthethenIUCNclassificationsand
evolution into separate sibling species over short guidelines.Theirrecommendationsweresubmittedtoa
distances, especially among insects with parasitic life- smallRDBselectionpanel,whichaimedforconsistency
styles (Elmes et al. 1999; Scho¨nrogge et al. 2002; across taxa and, with the agreement of each expert,
Thomas et al. in press). Too little is known about the certain species were reclassified. Nevertheless, as a
incidence of cryptic species or subspecific variation in member of that group, I and colleagues (Thomas &
insectsforthistobeconsideredforthe2010target,and Morris 1995) consider that several of the subsequent
hereafter I describe the monitoring of described, analyses that were made using the published list are
morphologically distinct species (morpho-species). invalidduetomajordifferencesinknowledgebetween
This is challenging enough: for example, most known thegroups(e.g.§5).
morpho-species of insect are tropical Coleoptera and, Despitetheirshortcomings,RDBsofinsectsmaybe
of those named to date, about 40% are known from a usefulinidentifyingthetypesofecosystem—andtypes
single site and many from a single specimen (Anon. of habitat within ecosystems—that most support
2003).This,withthefactthatprobably30–90%ofthe threatenedinsects(seefigure8),andtheyoftenprovide
world’sinsectspecieshaveyettobediscovered,means an invaluable foundation for setting conservation
that monitoring efforts inevitably depend much on priorities. They fall well short of providing a scientifi-
extrapolations from subsets of the better known cally rigorous process for reporting against the
groups, especially butterflies, and typically from 2010 target and, ideally, should represent the end
regions such as Europe, North America and Japan, product of other monitoring schemes rather than
where entomological interest is high. their foundation.
Phil.Trans.R.Soc.B(2005)
Monitoring changeand distribution of insects J.A. Thomas 341
3. ATLASES Table 1. Summary of recorder effort (total records received
(a) UK mapping schemes and mean number of records per species) of UK mapping
Forsomeinsecttaxaincertainnations,themappingof schemes for birds, plants and 39 groups of terrestrial or
species’ changing distributions has advanced usefully freshwaterinvertebrate.
from the dayswhen maps represented little more than
group totalrecords recordsperspecies
recorder distribution. The most comprehensive
examples provide sensitive measures not only of
bird 1252734 6023
changing distributions and range sizes (Parmesan
plant 9100000 5971
etal.1999)butcanbeusedassurrogatesforpopulation
butterfly(2001) 1857400 26534
censuses (Asher et al. 2001; Warren et al. 2001; butterfly(1984) 281424 4020
Thomas et al. 2004). Again, the UK system, overseen spider 451975 1773
by the Biological Records Centre (BRC), is the most macro-moth 378549 490
complete and longest-running in the world and its hoverfly 374784 1409
methodologies—which were derivedfrom the success- slug,snail 201240 887
fulplant(Perring&Walters1962;Prestonetal.2002) ground-beetle 141700 400
dragonfly 109117 1881
and bird projects (Sharrock 1976; Gibbons et al.
grasshopper 46346 786
1993)—have been widely adoptedin other nations.
milli-/centipede 34330 350
Reviews of the history of biological recording in
soldierbeetle 30412 422
Britain are given by Harding (1991a,b, 1995) and by
woodlice 27054 731
Burnett et al. (1995). Details of the 39 current UK brachycerafly 20980 141
recording schemes designed to map the changing lacewing 18508 226
distributions of invertebrate groups are found on the bumblebee 18505 685
BRC website (www.brc.ac.uk): they include annelids, longhornbeetle 12867 195
ticks, arachnids, woodlice, non-marine molluscs, harvestmen 11834 473
centipedes and millipedes, flatworms, beetles (16 atomariinebeetle 10814 226
water-flea 9559 111
groups),flies(18groups),mayflies,bugs,hymenoptera
ladybird 9340 212
(aculeates and sawflies only), lacewings and allies,
rovebeetle 8644 103
dragonflies, grasshoppers and allies, fleas, caddis flies
sepsidfly 6083 225
and butterflies and moths (table 1). Practical infor-
burnetmoth 5520 131
mation on each scheme includes contacts of local
picture-wingedfly 4921 76
organizersfordifferenttaxa,instructionsforrecording, leech 4414 276
standardized recording forms and procedures for crayfish 3896 649
quality control and the transfer and storage of data, muscidfly 2259 282
includingthe use of different softwareprogrammes. bee,wasp,ant 2054 93
flea 1939 52
cranefly 1894 271
(b) Case studyof two surveys of British
midge 1463 105
butterfly distributions
otherbeetles 1030 103
The two completed British butterfly mapping projects
ciidbeetles 736 147
differ from other schemes only in their intensity of
well/landshrimp 582 83
recordingandillustratetheusefulnessbothofmapping
water-slater 582 194
species’distributionstofinedetailandofrepeatingthis dolly/emphidfly 582 29
exercise after an interval, in this case, of two decades. flatworm 496 50
The methodologies and organization of each scheme parasiticwasps 431 144
aredescribedbyHeathetal.(1984)andbyAsheretal. tick(I.ricinus) 249 249
(2001) and are but briefly summarized here. Both fairyshrimp 163 83
projects aimed to achieve complete recording of the
presenceorabsenceofeverybutterflyspeciesineachof involving about 70 expert (mainly amateur) regional
the2861tenkilometre(10km!10km)gridsquaresof coordinatorsacrossthecountry.Recordingwasinitially
Britain. Field data were collected in 1970–1982 and undirected,butunder-recorded‘squares’weretargeted
1995–1999: the large majority of records were sub- in the second halfof each survey.
mitted by volunteer amateur recorders on a standar- Volunteerswereencouragedtorecordatthescaleof
dizedrecordcardforeachsitevisited.Thefirstsurvey the individual 1km!1km (1km) squares of the
was organized by BRC’s John Heath, who mobilized nationalgridorsmaller,butformostnorthernspecies,
entomologiststhroughoutthenationtosubmitrecords usefuldistributionmapswereachievedonlyatthescale
centrally or via regional organizers and local societies: of10kmsquares(figure1a).Inthisrespect,the10000
the second was organized by the Butterfly Conserva- recordersofthesecondsurveyachievedgreatersuccess
tion society (BC) through its network of regional than the ca 1500 recorders of the first survey, with
branches. After solving incompatibilities in software, 93% of records reported at 1km2 or finer resolutions
theyalsocaughtdatafromthenumerouslocalmapping compared with 51% submitted in 1970–1982; the
schemesestablishedinthe1980sand1990satthescale number of sightings of each species per visit was also
ofindividualcounties(Harding&Sheail1992).Inboth recordedin1995–1999(Asheretal.2001).Altogether,
surveys,therewasstrictqualitycontrolintheaccuracy thevolunteerrecorderssubmitted65826and437690
ofidentificationandrecordingofspeciesandlocalities, separate record cards (representing ca 2.1 million
Phil.Trans.R.Soc.B(2005)
342 J.A.Thomas Monitoringchange and distribution of insects
Figure 1. Monitoring change in the distribution and abundance of British butterflies. (a) Maps comparing results from
the1970–1982(Heathetal.1984)and1995–1999(Asheretal.2001)distributionsofBoloriaeuphrosyne,whichshowedthesixth
(10-percentile)greatestlossof10kmsquaresforaspeciesandPolygoniac-albumwhichexperiencedthefourth(7-percentile)
greatestincreasebetweensurveys.Timeseriesofmeanpopulationsizeofeachspeciesonsitesmonitoredoverthesameperiod
bytheBMSareshownbelow.(b)Thecorrelationbetween20yeartrendsinthepopulationindexesofthe27butterflyspecies
mostreliablymonitoredbytheBMSandchangesintheirrangesexpressedasthedifferenceintotaloccupancyof10kmsquares
byeachspeciesbetweenthetwomappingsurveys(fromWarrenetal.2001).
speciesrecords)inthetwosurveys,andsampled98.0% and/or southern boundaries, coinciding with warmer
and 99.6% of British 10km grid squares: more than summers, whereasonlyonespecies had shifted south.
90% of the missing 2% of squares in the first survey The second ‘Millenium’ atlas of British butterflies
were in biodiversity ‘cold-spots’ of Scotland within the (Asher et al. 2001) allowed the first quantitative
rangesofjustsixspecies(Thomasetal.2004). comparisonsofchangetobemadebetweentheresults
The first Atlas of Butterflies in Britain and Ireland of repeat surveys for an insect. Although the second
(Heath et al. 1984) achieved greater cover than (1995–1999)surveyachieveda6.6-foldhigherlevelof
had previously been thought possible for any taxon recordersandrecordingthaninthefirst(1970–1982),
of insects at the scale of a (228073km2) nation thenumberofrecordcardssubmittedperrecorderwas
(Ehrlich 1994). Although the older historical records the same (43.8 and 43.9 cards per recorder, respect-
were very patchy compared with the near-complete ively) and accumulation curves generated by sub-
1970–1982fieldsurvey,themapssuggestedthatabout sampling the second survey’s dataset revealed
70%ofUKbutterflyspecieshadexperiencedlossesas substantial redundancy (for analyses assuming 10km
measured by their occupancyof 10km squares during squareoccupancy)inthenumberofrecordssubmitted
the twentieth century (Thomas 1984, 1991), even formostregionsin1995–1999(table1).Toachievethe
though several species had—in some cases simul- most accurate comparisons between the two datasets,
taneously—experiencedsubstantialnorthwardsexpan- Warrenetal.(2001)employedasubsamplingroutineto
sions of their historical range boundaries in recent correct for the difference in recorder effort: the 1995–
decades (Pollard & Yates 1993). The role of climate 1999dataweresampled30timesbyrandomlyselecting
change as one of the drivers of species’ range changes the1970–1982 number ofrecord cardsfromthesame
wassuggestedbyParmesanetal.(1999),whoanalysed 1995–1999 data, subsampling separately for each
mapping schemes from the UK, Sweden, Finland, 100km Ordnance Survey grid square to retain the
Estonia, southeast France, Catalonia, Algeria, Tunisia broad geographical distribution pattern of the 1970–
and Morocco. They showed that, in a sample of 35 1982 recording effort. Results were robust when
non-migratory species, 63% had ranges that had reanalysedusingdifferentsubsamplingregimes,includ-
shifted 33–240km northwards along their northern ingratiosofrecordcardsper10kmsquare,thenumber
Phil.Trans.R.Soc.B(2005)
Monitoring changeand distribution of insects J.A. Thomas 343
of records rather than of cards submitted, and when outweighedpositiveresponsestoclimatewarming.
distributionchangewasmeasuredasthesquarerootin Only half of those species (e.g. Polygonia c-album,
area(insquarekilometres)rather thanasproportional figure 1a) that were both mobile and habitat
change. For 43 (74% of) species, subsampling results generalists increased their distribution sizes over
werefurthervalidatedbycomparisonwithindependent this period, consistent with a climate explanation;
measuresofspecies’populationchangesoverthisperiod whereastheotherlessmobilegeneralistsand89%
(Warrenetal.2001). of the habitat specialists (e.g. Boloria euphrosyne,
The existence of a comparable pair of datasets of figure 1a) declined in distribution size, consistent
species’ distributions after a 20 year interval suggests with habitat limitation. Warren et al. (2001)
that atlases can be used to monitor change in certain conclude that, in future, the dual forces of habitat
insects for the 2010 target. For example, Warren et al. modificationandclimatechangearelikelytocause
(2001) used differences in the number of 10km specialists to decline, leaving biological commu-
squares occupied by each species between the two nities increasingly depauperate and dominated by
surveys to measure the comparative loss or gain of mobile and widespreadhabitat generalists.
statusbydifferenttypesofspecies.AlthoughCowleyet
(ii) At the first Royal Society Discussion Meeting on
al. (1999) showed that this coarse-grained (10km
EstimatingExtinction Rates(Lawton&May1995),
scale) sampling may underemphasize more local
thekeyentomologicaldebatewaswhether‘recent’
changesinstatus,weusedthetimeseriesofpopulation
extinction rates among insect species matched the
change generated by the UK Butterfly Monitoring
well-recorded losses of other groups. On the one
Scheme(§4)forthesamespeciesoverthesameinterval
hand, the reported extinction rates of both
toshowthatrangechangesexpressedatthe10kmscale
historical and sub-fossil (Quaternary) species of
were correlated with trends in the mean size of
insect were 2–3 orders of magnitude lower than
individual populations of each butterfly (figure 1b
thoseknownfor vertebratesandcertainplanttaxa
fromWarrenetal.2001).Thisrelationshipispredicted
(Coope1994;Mayetal.1995);ontheotherhand,
in theory (Brown 1984; Gaston & Lawton 1988;
historical losses of butterflies and dragonflies,
Lawton 1995) but the closeness of the correlation
recorded at the scale of populations and meta-
between mapping and monitoring scheme results is
populations,weresignificantlyhigherthanthoseof
reassuring, and allows range changes (figure 1a) to be
vascular plants, birds, amphibia, reptiles and
consideredas asurrogate for abundance, making each
mammals in the same sites and landscapes
survey effectively a population census (Thomas et al.
(Thomas & Morris 1994). One hypothesis was
2004).
that insects had experienced comparable rates of
Asideeffectofmappingandmonitoringschemesis
extinction to higher taxa but, due to low levels of
that a growing number of volunteers grow to enjoy
early recording, the rarer species, which were also
recording in the field (Pollard & Yates 1993); indeed,
the most extinction-prone, had been overlooked
in one mapping project, the largest number of records
andweremissingfromthebase-linelists(Mayetal.
received arrived unsolicited for the year after the
1995; Thomas & Clarke 2004). The alternative
projecthadended(Thomasetal.1998b)!Givencentral
theory was that insect populations were more
institutional support and funding to scrutinize and
dynamic than vertebrates at the local level, and
processthedata(ca£150000isrequiredayeartorun
underwent rapid extinctions and colonizations of
UKinsectmappingschemes),itisthusacomparatively
individual sites and landscapes while persisting as
simpletasktoorganizerepeatsurveys:thechallenge is
species at alarger scale (Coope 1994).
in launching the first one. Thus, Butterfly Conserva-
tionhasembarkeduponrepeatsurveysofsimilarcover The publication of twin atlases not only of
to that achieved in 1995–1999 (Asher et al. 2001) for butterflies but also of changing bird and plant
2000–2004 and 2005–2009 (Anon. 2004; R. Fox, distributions, surveyed to the same 10km scale
personal communication). If successful, the sequence overasimilar periodinBritain,permittedthefirst
of four atlases should provide the rigorous process direct comparison to be made of regional extinc-
requiredforBritishbutterfliesforreportingagainstthe tions in these three groups across a (national)
2010 target. However, already the two completed landmassthatwas,inthiscase,aboutfourordersof
measurements of species’ distributions, when com- magnitude greater than the area occupied by a
bined with other datasets, provide insights about the typical meta-population of a British butterfly
drivers and sensitivity of insects to environmental (Thomas 1995; Thomas et al. 2004). We found
change. Twoexampleswill suffice: thatbutterfliesexperiencedgreater netlosses than
plants or birds at this scale, disappearing on
(i) Warren et al. (2001) evaluated change in the average from 13% of their previously occupied
distributionandabundancesof46butterflyspecies 10km squares (figure 2). We concluded that past
that reach their northern climatic range margins insect losses had been underestimated because
in Britain, where climate change and habitat the most threatened (rare) species had been over-
degradation are opposing forces. Although these looked in baseline lists (Thomas & Clarke 2004;
insects might be expected to have responded §5d), and, applying Pimm et al.’s (1995) ‘cookie-
positively to climate warming over the last 30 cutter model’, we tentatively suggested that, if
years (Roy et al. 2001), three-quarters of them insectselsewhereintheworldhadsimilarlevelsof
declined: negative responses to habitat loss sensitivity to the British butterflies, the known
Phil.Trans.R.Soc.B(2005)
344 J.A.Thomas Monitoringchange and distribution of insects
Europe, where several national atlases have been
(a)
25 published(vanSwaay&Warren2003),somematching
butterflies the recording effort and cover achieved by the first
20
s median change –13% British atlas. Good examples include atlases for
e
ci 15 Switzerland (Geiger 1987), the Netherlands (Tax
e
p
% of s 150 (1B9e8n9e)s,ˇ D&enKmoanrvkicˇ(kSatol2t0ze021)99a6n)d, tFhelanCdzeercsh(RMepaeusbl&ic
Van Dyck 1999). However, since none has achieved a
0 comparable repeat survey of distributions, these are
–100 –60 –20 20 60 100 140 180 >200
unlikely to achieve the objective of rigorously
30 measuring change by 2010.
s 25 birds Mapping of other invertebrates in Britain also lags
cie 20 median change –2% toofarbehindthatofbutterfliestofulfilthe2010target.
e
sp 15 Nevertheless, realistic baseline maps are possible for
% of 10 severalgroups(seefigure9),andthelevelofrecording
in some suggests that the required cover for useful
5
comparisonwilleventuallybeachievedthroughvolun-
0
–100 –60 –20 20 60 100 140 180 >200 teer efforts (table 1). If UK entomologists continue to
follow the lead of botanists, of whom only 1500
30 recorders made the 9 million records (ca 6000 per
s 25 plants species) submitted for the recent atlas of 1254 native
cie 20 median change +7% species of plant (Preston et al. 2002), recording levels
e
sp 15 adequate for quantitative analysis are surely attainable
of 10 for the smaller, more popular groups (table 1); for
%
5 example, more than 1500 recorders have already
submitted records of Odonata (H. Arnold, personal
0
–100 –60 –20 20 60 100 140 180 >200 communication).Nopreciseanalysishasbeenmadeof
% change in occupied 10 km squares the intensity of recording needed to make direct
comparisons between sequential mapping surveys of
(b)
the same group, but the cover achieved for birds and
100
plants of roughly twice as many records per species as
s the number of 10km squaresbeing surveyed (2861 in
e
ci 75 Britain)—or sixtimes theoccupancyofsquaresbythe
e
sp species of median range-size in each taxon—was
of adequate(see‘SupplementaryInformation’inThomas
% 50
e etal.2004)andis perhaps auseful guide (table 1). By
ativ thesecriteria,thefirstsurveyofbutterfliesfellshortby
ul 25 about 25%, the second butterfly survey captured five
m
u timesmorerecordsthanwererequired(for10kmscale
c
analyses) and those of dragonflies, spiders and hover-
0
flies need about two to three times their current cover
–100 –50 0 50 100 150
(table 1).
% change in occupied 10 km squares
Figure2.Changesinthenumberof10kmsquaresinBritain
4. TIME-SERIES AND ABSOLUTE POPULATION
occupiedbynativebutterfly,birdandplantspeciesbetween
thetwocensusesofeachtaxon(fromThomasetal.2004).(a) ESTIMATES
Frequency distributions: blackZextinct species, hatchedZ (a) Time-series
decliningspecies,dotsZincreasingspecies.Medianbutterfly Amore sensitiveapproachthan mapping distributions
species!median bird species (p!0.001)!median plant istomonitor trendsinpopulationsizesofspeciesover
native species (p!0.001). (b) Cumulative frequency distri-
time. At least three schemes exist in the UK that have
butions: solid line, butterflies; dotted line, birds; dashed
sampledannuallyatfixedsitesfor morethan25years,
line,plants.ButterflydeclinesObirddeclinesOplantdeclines
measuring aphids, macro-moths and butterflies. The
(p!0.001).
first, involving standard 12.2m high suction traps run
by the Biotechnology & Biological Sciences Research
global extinction rates of vertebrate and plant
Council’s Rothamsted Insect Survey since 1965, is
species have an unrecorded parallel among the
primarilydesignedtomonitorpestspeciesofaphid:its
invertebrates,strengtheningthehypothesisthatthe
usefulness to measure change in rarer species for
naturalworld maybeapproaching the sixth major
biodiversity conservation is constrained by the small
extinctioneventinitshistory(Thomasetal.2004).
numberofsamplingpoints,16inBritainandafurther
57in18otherEuropeancounties(Denholmetal.2001).
(c) Other insect mapping schemes The Rothamsted moth survey began in 1965 and
Many schemes for butterflies similar to that described annuallysamplesabout100sitesacrossBritainusinga
for Britain have been initiated elsewhere, especially in standard Rothamsted Light Trap (Denholm et al.
Phil.Trans.R.Soc.B(2005)
Monitoring changeand distribution of insects J.A. Thomas 345
2001). Staffed mainly by volunteers, it collates daily weather variation (Thomas et al. 1998a) and which
catchesofmacro-moths.Theassumptionthatvariation varied in density by up to three orders of magnitude
in the numbers trapped measures change in absolute betweensitesoronthesamesiteover thelonger term,
population size has been confirmed for only a few reflectingdifferencesorchangeinthequalityofsource
species (Raimondo et al. 2003) but is expected to be habitatfromoptimaltosuboptimalsites(e.g.figure4b;
valid.Thisschemeiscurrentlyusedmainlytomeasure Thomas 1983a,b, 1984). Crucially, we found
other aspects of Lepidoptera ecology, such as phenol- that simultaneous estimates of the absolute size of
ogical responses to climate change, for which it is populations measured by (time-consuming) MRR
sensitive (Sparks et al. 2001), although it has recently techniquesandtransect countswereclosely correlated
been extended to estimate rates of decline in some forthewiderangeofspeciestested:theexampleshown
widespread species (Conrad et al. in press). Similar in figure 4b tests an application of the technique to
schemesareruninIrelandandFranceandclearlyhave standardize counts from different sites during surveys
thepotentialtoprovidearigorousprocessforreporting (§4b; Thomas 1983a) but the same principle applies.
against the 2010 summit. Figure4balsoteststheassumptionthatindividualsofa
The third scheme, which monitors butterfly abun- speciesareequallyvisibleinpoorandoptimumhabitat:
dance, has been thoroughly tested and has been four sites are shown supporting M. cinxia populations
successfully applied to monitor other day-flying insect that vary in density by 20 to 30-fold on the peak day,
groups. It is therefore described in greater detail here. representing the full range in quality of source habitat
The method depends on standardized transect counts for this species in Britain; but two sites were also
ofadultbutterflies(Pollardetal.1975;Pollard&Yates sampled four times during the emergence, when the
1993) and is summarized in figure 3 in the simplified habitat may be considered constant but the density of
form used to instruct volunteer recorders. First, a adultswasverydifferent.Inthisandotherspecies,the
representativebiotope(Breretonetal.2003)isselected correlation between absolute population size and
for long-term monitoring and a fixed transect route is transect index is no closer for within-site comparisons
chosen, typically stratified into up to 15 sections to than it is betweensites.
subsamplemajor variationsinhabitatormanagement. Pollard et al.’s (1975) method measures relative
This route is walked at least once a week from 1 changesinpopulationsizeswithinanindividualspecies
April to 29 September (in the UK) under defined and is applicable to all the British butterfly species
conditions of weather, time of day, etc., when adult apart from the 2–3 inhabitants of woodland canopies.
butterflies are active, and every sighting of each Itisimportanttonotethattheabsoluteindexvaluesfor
species made in an imaginary 5m!5m!5m box is different species on the same site are not directly
counted in each section. At the end of the year, the comparable without correction for variation in inter-
mean weekly counts are summed for each species to specificdifferencesinbehaviourandhencevisibility.In
provide an index of abundance for each generation. addition,eachsample(transect)differsfromthebetter
Similar data from other sites are collated centrally known bird monitoring schemes in two respects. (i)
to produce a regional or national index of abun- Due to the relative immobility and small patch size of
dance for each species, and these in turn generate a about 75% of butterfly species (Thomas 1984), each
time-series of population change when counts are transect samples one (or more) entire ‘closed’ popu-
repeated on the same sites in successive years lations of most species in the biotope, rather than a
(figures 1a, 5 and 6). small part of an extended or open population. (ii)
Pollard et al.’s (1975) method was not the first British butterflies are short-lived with one or two
attempttousetransectcountstomonitorabundancein emergencesayear;inmostspecies,annualindexesreflect
butterflies (e.g. Moore 1975), but it differed in that changesinnumbersbetweendiscretegenerations.
strict standardized criteria were defined for the AnationalBMSwaslaunchedin1976,organizedby
conditions under which recording was permitted and Ernie Pollard at ITE’s (now CEH) Monks Wood
becausetheseandotherassumptionsweretestedinthe laboratory. Twenty-eight years later (Z28 or 56
field. Full details of methodology, the launch of a generations of each butterfly species), it coordinates
national Butterfly Monitoring Scheme (BMS), the the records of 134 long-term sites distributed quite
treatment of data from sites with very large or very evenlyacrosstheUK,butwithasmallbiastothesouth
smallnumbersofaspeciestoavoidbiasinthecollated where butterfly species-richness is greatest (figure 9).
index,theincorporationofnewsitesorthosethatdrop As already noted, butterfly monitoring became so
outofthescheme,thedetectionofsignificanttrendsin popular—and the results so clearly useful—that many
species abundance and many of the tests used to independentrecordersbegantheirown‘Pollardwalks’
validate the method are described in detail by Pollard (as they are affectionately known), sampling sites of
et al. (1975), Pollard (1977, 1979), Thomas (1983a), local interest. The surplus data were beyond the
Pollard & Yates (1993), Pollard et al. (1995) and resourcesoftheBMSdatamanagementandunnecess-
Moss&Pollard(1993).Onlyafewkeyassumptionsare ary to achieve that scheme’s aims, but a separate
consideredhere(figure4).Forexample,wefoundthat booklet, Instructions for Independent Recorders, was
recorder effects were negligible (e.g. figure 4a, Pollard written to achieve the same rigour of recording as in
et al. 1975) in comparison to the size of differences the BMS. In the 1990s, the burgeoning number of
being compared within species, which typically independent recorders was coordinated by BC and
fluctuated by up to one order of magnitude from extended into an equally professional parallel
generation to generation on the same site due to monitoring scheme that sampled ca 450 sites in 2003
Phil.Trans.R.Soc.B(2005)
346 J.A.Thomas Monitoringchange and distribution of insects
1984, 1991). In addition, BC is now establishing
additional transects on more intensive ‘ordinary’
fixed transect route
farmlandinorderbettertomonitorsuccessinmeeting
chosen to sample
2010 targets in the UK as a whole (Brereton et al.
biotope 2003). However, solving the current bias is less
straightforward than was the case for breeding birds,
for two reasons. (i) Because most butterflies live in
small populations restricted to discrete patches of
suitable habitat, intensive farmland is today almost
bereft ofbutterflies toanextentthatdoesnotapplyto
weekly counts birds, and the few species that are seen there are
common and well recorded elsewhere (although the
strict criteria for
counting, time of trendsmaydiffer),or thosewithopenpopulations.(ii)
day, weather, etc. It is less easy to persuade volunteers to make weekly
transects,yearafter year,onlandwheretheyseeafew
individuals of afew common species.
Despite these problems, both monitoring schemes
already contribute strongly to a scientifically rigorous
processforreportingagainstthe2010target,especially
for the scarcer species that are largely confined to
results: weekly counts of
en protected land. They also provide insights about the
each species present e
no. s defrfievcetrssaonfdpcolpimulaatteiocnhacnhgaeng(ee,.gr.anPgoilnlagrdfroetmalw.e1a9th9e7r;
Roy et al. 2001) to habitat degradation (e.g. Thomas
1 2 3 4 5 6 7 8 9
1983b). Many other examples are summarized by
week
Annual/generation index Pollard & Yates (1993). More relevant here is how
for each species easily can trends be detected in species or groups of
Σ weekly counts e.g. speciesusingtransectindices?First,itshouldbenoted
1+3+9+12+7+3+3+2 = 40 thatthepopulationsofmostbutterflyspeciesfluctuate
considerably between generations, often synchronized
multi-site population
over short to medium distances (Pollard 1991), in
trends for each species x
de response to variation in weather (Pollard 1988;
n
i Roy et al. 2001). As theory predicts, fluctuations are
year amplifiedtowardstheclimaticedgesofspecies’ranges,
making time series of longer than 15 years essential if
Figure3.Diagramoffivestepsinnationalschemesforusing underlyingtrendsaretobedetected(figure5,Thomas
transectcountstogeneratetime-seriesofbutterflypopulation etal.1994;Pollardetal.1995).The28yearand19year
changes (from T. Brereton, personal communication); see
runs of the BMS and BC schemes are thus adequate.
textfordetails.
For example, the collated trends in specialist and
generalistspeciesofbutterfly(seePollard&Eversham
(www.butterfly-conservation.org). Like the BMS, BC
1995;Warrenetal.2001fordefinitions)fromtheBMS
generateannual indexesof abundanceand established
show a marked divergence since 1976 (figure 6a),
anetworkoflocalcoordinatorstoensurequalitycontrol;
confirming a key result obtained from comparisons of
BC also developed TRANSECT WALKER, a user-friendly
the two butterfly atlases (Warren et al. 2001). In
software package that can be downloaded free from
addition, grassland species showed an increasing
the BC website, which enables individual recorders to
population trend over 28 years, whereas woodland
generateindicesandgraphsfortheirownsites.
species declined (figure 6b), although note that most
The BC scheme has shorter time-series than the
grasslandsitesweresemi-naturalfarmedbiotopessuch
BMS (useful indexes start from ca 1985: figure 6c),
asunimprovedcalcareousdownland.Althoughwarmer
samples four times the number of populations and is
weather has played a role (Thomas, Bodsworth et al.
stratified to sample a higher proportion of southern
2001; Warren et al. 2001), for many grassland species
sites: it therefore generates more reliable time-series
this encouraging increase results from targeted
for certain rare species due to the larger number of
conservationmanagementfollowingknowledgegained
populations sampled. However, apart from the wood- of the ecological requirements of many species in the
land sites, both schemes are currently biased in that 1970s and 1980s (reviews: Thomas 1983a,b, 1991).
about a third of transects sample biotopes that have Someextensiveagri-environmentalschemeshavebeen
some form of environmental protection, such as land particularlysuccessful:theincreaseofLysandracoridon,
in agri-environmental schemes, nature reserves, Sites an indicator of high quality calcareous grassland, on
of Special Scientific Interest (SSSIs) and national targeted compared with other sites demonstrates the
parks.Itisimportantthat‘protected’sitesremainwell usefulness of the BC scheme for monitoring change
recorded, for they support all the surviving popu- (figure 6c).
lations of certain UK butterfly species and the In 2005, the two UK monitoring schemes will be
majority of populations of several others (Thomas integrated (subject tofunding), then merged, and that
Phil.Trans.R.Soc.B(2005)
Monitoring changeand distribution of insects J.A. Thomas 347
bumblebees, hoverflies, dragonflies (R. Cox, personal
(a)
communication)andparasiticHymenoptera(Thomas,
120
Pieris napi unpublished) appear promising but, apart from the
nt Hymenoptera, have yet to be tested against indepen-
ou 80
c dent measures of population size.
ct
e
ns 40
a (b) Surveysofabsolute population size
tr
Thebutterflytransectmethod,whichrecordschanging
0 •
densities of species over time, was adapted (Thomas
1983a)tomeasuretheabsolutesizesofpopulationson
120 differentsitesinasingleyear.Theaimwastoprovidea
Aphantopus hyperantus
nt simple method of identifying the total number,
u
o 80 boundaries and size of every population of a scarce
c
ct (especiallyBAP)speciesintheUK,asawholeorovera
e
ns 40 substantial part of its range, from a one-off survey in
tra • which most siteswerevisited onlyonce (e.g. figure 7).
0 The method (Thomas 1983a) differs from BMS
May Jun e July Aug Sept transects in four ways: the transect route takes a
stratified sample of each population being measured,
(b)
accountistakenofthelengthoftransectandthesizeof
ct patch occupied by each population, and the new
se 2.0 population size indexes for each species are calibrated
n
a melitaea cinxia
m tr r 2 = 99%; p < 0.001 ftreocmhnaiqbuseosluotenmaefaeswurreesfeorfenncuemsbiteerss(mfiagduereu4sibn)g. MI aRlsRo
o
fr 1.5 suggested a formula for adjusting for the dates in the
d
ate emergence when different sites were sampled, again
ul using reference sites. This later proved unreliable and
alc 1.0 all surveys are now restricted to the peak two weeks
c
p) ofaspecies’emergence.Consequently,theresultsfrom
x ( each site can be assigned to only one of 4–5 broad
de 0.5
n classesofpopulationsizes(figure7),alevelofprecision
n i that is nevertheless adequate to satisfy most conserva-
o
ati tion and scientific needs. This method is less intrusive
ul 0
p thattraditionalMRRtechniquesofestimatingabsolute
o 1 10 100 1000 3000
p numbersandtakesaboutone-fifththeamountoftime
actual population size ± se estimated by MRR on 4 sites
per population (Thomas 1983a).
Figure 4. Testing assumptions in transect recording of Surveys encompassing all or the majority of a
butterflies. (a) Recorder effect: results of three different species’ range have been made using this technique
recorders (symbols) making repeated counts of two species onmorethan25%oftheUKbutterflyspeciestodate,
along the same transect route during the season (from
with repeat surveys being made at intervals of 10–20
Pollardetal.1975).(b)Comparisonoftransectcountswith
years for six species. In future, BC aims to repeat full
absolute population estimates of one species made on the
surveysofkeydeclining(BAP)speciesevery3–5years
same days on four sites (symbols) at different times in the
(N. A. D. Bourn, personal communication). In
emergence(fromThomas1983a).
addition to providing a near-complete description of a
species’ national status in a particular year, the results
part run by BC will expand to better monitor success
provided insights into the population structures of
in meeting government targets concerning the UK
differentspeciesandshowedthatasinglebiotopeisland
Biodiversity Action Plan (BAP), PSA (SSSI), agri-
may support several demographically distinct popu-
environmental schemes and sustainable development
lations of the same species (Thomas 1983b, 1991;
(T. Brereton, personal communication). Each scheme
Thomas et al. 1999a,b). This, in turn, enabled the
representsoutstandingvaluethankstotheca15000h
distribution of individual populations within several
free labour provided annually by ca 1500 skilled
meta-populationtobeidentifiedwithgreaterprecision
volunteer recorders; however, professional coord-
than previously, and was useful in assessing the
ination is essential and the full cost of developing and
importance of meta-population dynamics as a driver
running both schemes to meet future targets over the
of observed species’ changes (Thomas et al. 1992;
next 3 years is ca £200000 a year (D. Roy &
Thomaset al.2001b).
T. Brereton,personal communication).
SuccessfulBMSsusingthesamemethodologyhave
been established in the Netherlands (van Swaay et al. 5. HOWREPRESENTATIVE ARE BUTTERFLIES
2002), Flanders, Finland and Catalonia, and uncoor- OFOTHERINSECTS?
dinatedannualtransectrecordingispractisedonmany (a) Introduction
sites outside Europe, for example in Japan and the Despite the success of Brown’s (1991, 1996a,b)
USA. Small-scale attempts to extend the method to pioneeringuseofawiderrangeofgroupsasindicators
Phil.Trans.R.Soc.B(2005)
348 J.A.Thomas Monitoringchange and distribution of insects
200
Southern (core) pops
e 115 (a)
ance Northern (edge) pops ±e s 110
nd160 nc 105 generalist spp
ndex of abu120 x of abunda 1990050
collated i 80 collated inde 77880505 specialist spp
1975 1980 1985 1990 1995 2000 2005
40 year
1977 1980 1984 1988 1992
140 se 110 (b)
±
e grassland spp
c
n
da 100
n
u
V ab
pulation C 100 d index of 9800 woodland spp
o e
p at
60 oll
c 70
1975 1980 1985 1990 1995 2000 2005
year
1000
20 (c)
0 200 400 600 800 e)
distance north (km) al
c
s
g
Figure5.Fluctuationsinthesizeofmonitoredpopulationsof o
(l
butterfliesatdifferentpointsintheirranges.(a)Thefivemost x 100
e
northern (edge of range) sites of Maniola jurtina compared d
n
withthefivemostsoutherly(core)populations.(b)CVsofall d i
monitored populations of M. jurtina plotted against latitude ate
(fromThomasetal.1994). coll all sites scheme non - scheme
10
of biodiversity-richness and change (§5d), butterflies 1985 1987 1989 1991 1993 1995 1997 1999
arecurrentlytheonlyinsecttaxonthatitispracticalto
Figure 6. Application of trends in butterfly time series as
monitorwithprecisioninmanypartsoftheworld(e.g.
a measure of 2010 criteria. (a) Variation in trends of
Brown & Brown 1992; Ehrlich 1994). Since, for the generalist (bZC0.30, r2Z14%, p!0.05) and specialist
foreseeablefuture,theyarelikelytostayelevatedtothis species (bZK0.34, r2Z27%, p!0.005) measured over 28
role of ‘miner’s canary’ (Attenborough 2001) or years(Z28or56generations;seePollard&Eversham1995;
‘honorary birds’ (May et al. 1995) vis a` vis insect Warren et al. 2001 for definitions). (b) Variation in
extinction rates, it is important to examine how well grassland (bZC0.32, r2Z18%, p!0.03) and woodland
species (bZK2.0, r2Z9%, ns). (c) Trends in Lysandra
butterfliesrepresent change in other insects.
coridon on 68 sites managed under agri-environmental
Most experts conclude that butterflies are
schemes and 67 sites not in the scheme (managedOnon-
reasonable, albeit imperfect, representatives of other
managed, p!0.001). (a),(b) from CEH’s BMS; (c) from
invertebrates (Ehrlich 1994; Master et al. 2000). In
the Butterfly Conservation scheme (T. Brereton, personal
contrast, Hambler & Speight (1996, 2004) use the communication).
listed extinctions of all taxa in the British RDB (Shirt
1987) to argue that butterflies have experienced moresensitivethanotherinvertebratestoenvironmen-
amplified losses compared with other insects and that talchangedue to their life-stylesor to thermophily.
their useasindicatorsofgeneralchangeisinappropri-
ate, because ‘butterflies (being mostly warmth-loving (b) Butterfly lifestyles
and herbivorous) are atypical invertebrates that are Being terrestrial in all stages of their life-cycle,
relativelysensitivetoclimaticfluctuationsandthusgive butterflies are clearly inappropriate indicators of
a potentially misleading guide to extinction rates and freshwater species, whose declines are believed to be
human impacts’. In §5d, I suggest that the first substantiallygreater thanthoseofterrestrialones(e.g.
argumentisflawedduetoanartefactofpastrecording Shirt 1987; Master et al. 2000). They do, however,
levels. First, I consider whether British butterflies are inhabit all the main terrestrial ecosystems, at least of
Phil.Trans.R.Soc.B(2005)