Lichenologist 33(3): 249-260 (2001)doi:10.1006/lich.2000.0320, available online at http://www.idealibrary.com on IDEJ^ L'

Evaluation of bio-indication methods using epiphytes bycorrelating with SO2-pollution parameters

Wouter GEEBELEN and Maurice HOFFMANN

Abstract: In order to evaluate the relative usefulness of the numerous bio-indication methods thatuse epiphytes as indicators for air pollution, a range of these different methods was applied within thesame region (Flanders, Northern Belgium). The study area shows important air pollution gradientsand little climatic differentiation. Ninety-one IAP indices were calculated with data collected in theimmediate vicinity of SO2 measuring stations on one randomly chosen tree, on ten neighbouring treesand in a sampling grid placed on the tree carrying the largest number of epiphytes. A large majorityof indices showed a significant Spearman rank order correlation with mean SO2 winter values over the5 years 1990-94. Indices based on a selection of 17 lichens gave better results than those based onall lichens. The latter indices showed higher correlation than those using all epiphytes (includingbryophytes). Indices using a quantitative estimate of cover or abundance or a combination of thesedid not result in a higher correlation. Sampling ten trees gave slightly higher correlation than the gridmethod or sampling of one tree. Taking into account supplementary evaluation criteria (timeinvestment and degree of lichenological knowledge required), sampling of 17 bio-indicator speciesusing a sampling grid is recommended within the study area. Monitoring the presence of allcorticolous lichens within a sampling grid is recommended in general.

:f) 2001 The British Lichen Society

IntroductionSince corticolous lichens react strongly anddifferentially to acidifying air pollution, theyare extremely useful as bio-indicators. SinceNylander's (1886) lecture on the povertyof the lichen flora in the centre of Paris,innumerable accounts have been publishedon the relation between the vicinity of in-dustrial and urbanized places and SO2-concentration on the one hand, and thepresence of corticolous lichens on the other.

W. Geebelen and M. Hoffmann: University of Ghent,Dept. Biology, K. L. Ledeganckstraat 35, B-9000Ghent, Belgium.W. Geebelen. Present address: Limburg UniversityCentre, Centre for Environmental Sciences, Universi-taire Campus, B-3590 Diepenbeek, Belgium.M. Hoffmann. Present address: Institute of NatureConservation, Kliniekstraat 25, B-1070 Brussels,Belgium.

However, bio-indication methods differstrongly, using parameters such as numberof species, a large spectrum of poleophobyindices, presence, abundance and/or coverof all or individual species, etc. Differentmethods can result in different diagnoses forthe same region. Other studies attempting todevelop a generally applicable method are,among others, Kirschbaum & Windisch(1995), Nimis et al. (1989), Van Haluwyn& Lerond (1988). Comparison of data onepiphytes and air pollution from the sameregion was also performed by Herzig et al.(1985) and Herzig & Urech (1991).

To develop a generally applicable method,we compared as many as possible bio-indication methods based on epiphyteswithin the same region and using the samesampling sites. The result of the studyshould be a standard method selectedon the basis of reliability and practical

0024-2829/01/030249+12 835.00/0 2001 The British Lichen Society

250 THE LICHENOLOGIST Vol. 33

considerations. The methods were judgedon (1) Spearman rank order correlation withactual SO2 data, (2) time investment, and(3) degree of lichenological knowledgerequired.

Materials and MethodsStudy area

Flanders (Northern Belgium) was chosen as a studyarea. Earlier research using epiphytes as indicators ofair pollution in the area (Barkman 1963; Iserentant& Margot 1963; De Sloover & Lambinon 1965;Caekebeke 1986; Quanten 1986; Van der Gucht &Hoffmann 1990; Hoffmann 1993; Tanghe & Richel1996) proved the usefulness of bio-indication methods.A relatively dense S02-monitoring system is present(Fig. 1); steep gradients of air pollution, ranging fromrather moderately polluted (yearly average 1994: 8 ugSO2 m~3) to heavily polluted areas (yearly average1994: 53 ug SO2 m ~ 3) are present (yearly average of allstations in 1994: 23 ug SO2 m^3). Flanders covers anarea of approximately 13 500 km2. The area has anextremely high population density (430 inh/km2) andseveral large conurbations (Brussels, Antwerp, Ghent)where intense economic and industrial activities areconcentrated. Recently the overall air pollution situ-ation has improved considerably. During the period1950-1990, the area suffered from much greater airpollution levels than during the last decade. Flandershas a temperate oceanic climate with a slightly increas-ing continentality to the east. Mean annual temperatureranges between 9-3 and 10-6C (average 9-8C) andannual rainfall between 670 and 870 mm (average770 mm) (Walther & Lieth 1967).

SO2 concentration data were obtained from theformer Institute of Hygiene and Epidemiology(IHE, 1983-1993a, fc, 199O-1992) and the VlaamseMilieumaatschappij (1993, 19931994). In the studyarea, 91 SO2-measuring stations are present. To evalu-ate the indicator value of the biological data, they wereinitially compared with yearly mean values, wintermean values and 98-percentile values of atmosphericSO2 pollution for one to five years. Calculations pre-sented here are based on the comparison with meanSO2 winter concentrations over the 5 years 1990-1994.We used a SO2 parameter based on a relatively longperiod of 5 years instead of values of the year ofcollecting biological data, since it is generally acceptedthat epiphytes are indicators of the general air pollutionsituation rather then indicators of the instantaneous airpollution level (see e.g. Herzig & Urech 1991). To beincluded in the calculation of the overall mean for thelast five years, at least three out of five years should havea reliable measurement value. In this way 57 SO2-measuring stations couJd be included m tht caJcu-

lations. We chose to use the SO2 mean winter concen-trations (October-March) because SO2 emission isgenerally higher in this period, as is the photosyntheticactivity of cryptogamic epiphytes (Stalfelt 1939;Barkman 1958; Kershaw 1985), which makes themmore vulnerable in this period.

In contrast to the well-distributed SO2-measuringnetwork, the presence of stations measuring other majorpollutants (NO, NO2, O3) is relatively poor in the studyarea. For this reason they were not considered in furtheranalysis of the data.

SamplingSampling was carried out in the autumn of 1994.

Within a radius of 1 km around each SO2-measuringstation, a well-exposed line of at least 10 trees wasselected from aerial photos. Sampled phorophytesshould have a well-grooved bark and a minimal per-imeter of 1 metre at a height of 1-2 metres. Preferencewas given to Populus x canadensis because of its generalepiphyte richness and its widespread presence in thearea (Hoffmann 1993). All 57 pre-selected SO2-measuring stations met these criteria (Fig. 1). Fromeach line of trees, one tree was randomly selected. Thistree and nine successive trees were used for a quantita-tive inventory of the epiphytes. They were sampled insuch a way that practically all known biomonitoringmethods could be covered.

Three sampling procedures were used at every site.The first procedure considers all epiphytes growing onone randomly chosen tree trunk from the base up to 2 mhigh. Frequency (A) and cover (B) of every lichenspecies was noted (Table 1). The second procedureincludes the first tree and nine successive trees on whichall epiphytes were identified. On every tree frequency(A) and cover (B) of every lichen species was noted.This procedure is similar to those used by Gilbert(1965), Leblanc & De Sloover (1970), Deruelle (1978),Van der Gucht & Hoffmann (1990) and Oksanen et al.(1991). They all sampled approximately 10 trees persite. Only Hawksworth & Rose (1970) sampled as manytrees as possible.

l"he third procedure was used only on trie treecarrying the largest number of species. A grid(100 x 40 cm, divided into 40 plots of 10 x 10 cm) wasplaced on the most luxuriantly overgrown side of thetree (almost without exception, the SW-side of thetree). The base of the grid was at 80 cm above groundlevel. The grid was kept in place by means of 4 nails.Species frequency was defined as the number of plots inwhich the species was present. Cover was estimatedusing the Londo scale (Londo 1984). Total epiphytecover was estimated as a percentage. The methodsdescribed by Britt (1987), Kirschbaum (1995),Liebendorfer et al. (1988), Nimis etal. (1990), Oksanenet al. (1991) and Ruoss (1992) could be coveredusing this procedure. They all used a sampling gridmethod; but the size and number of the units couldvary according to the author. This procedure wasfurther used to identify and characterize the epiphytewgtmion synmonomhfy,

N f 300

30K

ilom

eter

s

60

to o o tn 5 3 O > ex 5' B' a o n s- o P. IFI

G.

1.

Map

of S0

2-po

llutio

n

mea

surin

g st

atio

ns in

th

e st

udy

area

(Flan

ders,

N

orth

ern

Belg

ium

) sho

wing

th

e m

ean

SO2

win

ter

depo

sitio

nco

nce

ntra

tions

over

the

5 ye

ars

1990

-199

4;

shad

ed

area

s ar

e m

ajor

conurb

atio

ns.

to

252 THE LICHENOLOGIST Vol. 33

TABLE 1. Frequency and cover classes used to quantify thepresence of individual species on every tree

Class

The time needed for data collection on the first tree,on all 10 trees and on the grid was noted.

A Frequency(individuals)

1-22-55-10

10-30>30

B Cover(% of the bark)

1-20%20-40%40-60%60-80%80-100%

In the 1950s Barkman (1958) assigned a poleophobyindex (PB) to every species based on his field experiencewith those particular species in the Netherlands.Hoffmann (1993) assigned a poleophoby index (PH)based on distribution patterns of species in relation toSO2-deposition: species which increase in presence withincreasing emission were given a negative value, specieswhich decrease were given a positive value. The PH indexwas based on observations in East- and West-Flanders(part of the present study area). In a similar way, Wirth(1991) assigned a toxitolerance index (Tw), that has ahigher value for species with a higher tolerance to airpollution (which is the opposite to PB and PH).

The poleophoby indices of Barkman (1958) andHoffmann (1993) and the toxitolerance index of Wirth(1991) were used in the formulae of Herben & Liska(1986). All indices are listed in Table 2.

Indices of Atmospheric Purity (IAPs)Index of Atmospheric Purity values based on data

from one tree, 10 trees and grid sampling, were calcu-lated in as many ways as could be found in theliterature. Four major IAP-groups are distinguished.

IAPs based on number of speciesThe first group of IAPs simply takes into consider-

ation the total number of epiphytes or the total numberof corticolous lichens (Jones 1952; Barkman 1963;Gilbert 1965; De Wit 1976; Sergio & Bento-Pereira1981; De Bakker 1984; Van der Gucht & Hoffmann1990). According to this approach, IAP is defined as:IAP I. 1 = 2 lichens (van Dobben 1990; Hoffmann

1993; van Herk 1993)IAP I. 2 = S epiphytes (Hoffmann 1993)

IAPs using a poleophoby indexA second group of methods uses so-called air pollu-

tion sensitivity indices of species. Many authors use apoleophoby index Q, being the average number ofaccompanying species for every species. For all corti-colous lichen species present, a Qrvalue (Q,= averagenumber of accompanying lichens) was calculated and aQa-value was calculated (Qo = average number of ac-companying epiphytes). With application of Q r andQa-values (in formulae indicated as Q,), the followingformulae could be calculated using all epiphytes or onlycorticolous lichens (with n = the number of speciespresent):

IAP IV. 6 = -

IAPIV. 8 = -

IAPs using estimates of abundance and/or coverA third group of IAPs includes a quantitative estimate

of either cover (Jones 1952; Fenton 1960; Griffith1966; Trass 1968, 1973; Oksanen et al. 1991; Ruoss etal. 1991), or abundance (Britt 1987; Nimis et al. 1990;Aptroot & Roos 1993; Hoffmann 1993; Geebelen1995) or a combination of both (Iserentant & Margot1963; Leblanc & De Sloover 1970; Skorepa & Vitt1976; Liebendorfer et al. 1988).

Three formulae use an estimate of abundance (fre-quency). Leblanc & De Sloover (1970) use the index fi;combining cover and frequency. In Britt (1987) andNimis et al. (1990) F, is the frequency of all speciespresent in a sampling grid.

Frequency was quantified here on a five-point scale(Table 1), when calculations were made for one tree.When data for 10 trees were used, the number of treeswhere the species was present was used (1-10). Allthree formulae were calculated with Qrand Qa-values.When a sampling grid was used (in all three formulae),frequency was the number of plots where the specieswas found, converted to a tenfold scale.

IAP II. 1 & 2 = raQ, (De Sloover & Leblanc 1968)

IAPV.

IAP VI. 1 & 2 = -

(Leblanc & De Sloover 1970)

x/0 (Geebelen 1995)

IAP III. 1 & 2 = (Herben & Liska 1986) I A P VII. 1 & 2 = I F , (Britt 1987; Nimis et al. 1990)

IAPIV. 1 & 2 = - Q , .Trass (1968) attributes a toxic sensitivity index (a,-) to

(Herben & Liska 1986) species i based on the observation of the species

2001 Evaluation of bio-indication methodsGeebelen & Hoffmann 253TABLE 2. Epiphytes and their poleophoby index according to Barkman (PB) and Hoffmann (PH) and toxitolerance index

ofWirth (7V)

Desmococcus olivaceusLecanora conizaeoidesLepraria incanaLecanora expallensBuellia punctataLecanora chlaroteraDiploicia canescensPhyscia tenellaPhysconia griseaXanthoria candelariaPhaeophyscia orbicularisPhyscia adscendensPunctelia subrudectaParmelia sulcataCandelariella vitellinaXanthoria parietinaTrentepohlia umbrinaHypogymnia physodesLecidella elaeochromaOrthotrichum diaphanumCandelariella xanthostigmaLecanora carpineaBuellia griseovirensCalicium virideRinodina oleaeHysterium angustalum

122233444555555666

9999688457868

7

86

65545

- 1

21352

3

32424

4434

Xanthoria polycarpaPhyscia caesiaRamalina farinaceaPleurosticta acetabulumHypogymnia tubulosaParmelina tiliaceaPhysconia distortaCandelaria concolorMicarea prasinaPhyscia stellarisEvemia prunastriRamalina fastigiataHypotrachyna revolutaFlavoparmelia caperataMelanelia glabratulaRamalina fraxineaLecania cyrtellaHyperphyscia adglutinataOpegrapha atraAnaptychia ciliarisPseudevernia furfuraceaParmotrema chinenseA4elanelia exasperatulaParmelia saxatilisPlatismatia glauca

7777

8

8888

9

9999

1010

6665

44462

36233

1267

34

255555

TABLE 3. Transformation from estimated cover to tenfoldscale according to Trass (1968, 1973)

12345678910

1-3%3-5%5-10%

10-20%20-30%30-40%40-50%50-65%65-80%80-100%

a sampling grid. Therefore this formula was calculatedonly when sampling was done by means of a grid.

LAP VIII = V - ^ ^ - (Trass 1968, 1973)

The so-called 'lichen index' of Ruoss et al. (1991)was based on grid sampling, in which total coverand number of species were considered importantbio-indicator parameters:

IAPIX = DxAZ100(Ruoss et al. 1999; Ruoss 1992)

where D is total cover (in percent) and AZ (Artenzahl)is the number of species present.

in relation to air pollution. Relative cover of eachspecies was given by c,, total cover was estimated as apercentage (C,).

The toxic sensitivity a, was replaced here by thepoleophoby index (PH) of Hoffmann (1993), relativecover (c,) equals the re-scaled Londo cover estimates.The total cover estimated here as percentage cover wasconverted into the tenfold scale of Trass (1968, 1973;Table 3) and used as C,. Trass (1968, 1973) used

IAPs based on a selection of speciesA fourth group of IAPs only uses a selection of species

(bio-indicator species: BI). Hoffmann (1993) selected17 species for the western part of the present study area.They were denned as eurytopic species (species with awide general ecological amplitude), but with strictconstraints as far as acidifying air pollution is con-cerned. They were in order of decreasing toxitolerance:

254 THE LICHENOLOGIST Vol. 33

Lecanora conizaeoides, Buellia punctata, Parmelia sulcata,Lecanora expallens, Physcia tenella, Xanthoria parietina,Lecanora chlarotera, Ramalina farinacea, Puncteliasubrudecta, Lecidella elaeochroma, Pleurosticta acetabulum,Lecanora carpinea, Flavoparmelia caperata, Diploiciacanescens, Melanelia glabratula, Ramalina fastigiata andRamalina fraxinea (nomenclature follows Diederich &Serusiaux 2000).

Formulae cited earlier were also calculated usingthese 17 species instead of all epiphytes or all lichens:IAP 1.3, IAP II.3, IAP III.3, IAP IV.3, IAP IV.5, IAPIV.7, IAP V.3, IAP VI.3, IAP VII.2. IAP VIII.2 andIAP IX.2. Poleophoby indices (Q;, Qa) used in theseformulae were replaced by the average number ofaccompanying bio-indicator species (QBI). In addition,these formulae were also calculated with the poleo-phoby index (PH) of Hoffmann (1993) in IAP II.4, IAPIII.4, IAP IV.9, IAP V.4 and IAP VIA

In their qualitative bio-indication methodHawksworth & Rose (1970) make a distinction betweentrees with eutrophic and non-eutrophic bark. For eachthey distinguish different zones and with increasingSO2-concentration a number of species which can bepresent. When applying their approach, the specieswhich are entirely absent from Flanders (due to eco-logical or biogeographical reasons) were not considered.As a quantitative IAP based on their approach, wesummed the species present in the Hawksworth &Rose tables (indicated below as H&R tables) for botheutrophic and non-eutrophic bark.

IAP X = (species present in H&R tables)(Hawksworth & Rose 1970)

Thirty-five IAP formulae could be calculated and mostcould be calculated using data from three differentsampling procedures yielding a total of 91 different IAPvalues per sampled plot.

Bio-indication methods that could not be usedSome methods use the incompleteness of an epiphyte

community as an indication of air pollution (Barkman1963; Van Haluwyn & Lerond 1988; Wirth 1988).However, bio-indication methods using phytosociologi-cal data were eventually not used in this study, althoughour grid samples were analysed with Twinspan (Hill1979), enabling the identification of syntaxonomicalaffinities. Epiphyte vegetation showed affinities towardsthe Parmelion perlatae and Xanthorion parietinae. Manyof the releves syntaxonomically had to be classified asLecanoretum conizaeoidis, a syntaxon appearing only inair-polluted circumstances. Since it is not possible todetermine the original syntaxonomical association thatwould appear under less polluted circumstances, thesesites could not be tested for their incompleteness.Phytosociological methods were therefore not applied.

Methods using the presence of individual speciesfor determining epiphyte deserts, struggle zones andnormal zones (Sernander 1926; Showman 1975; DeBakker 1984; Hoffmann 1993) were not applied here,

because systematic sampling is required. This was notpossible, since we collected data around the non-randomly distributed SO2-measuring stations. Methodsusing characteristics of one species (Haugsja 1930;Showman 1988; Ackermann 1993) in relation to indus-trial activity or large agglomerations were not used inthis study. It would not be very useful to work out aprocedure that is dependent on the behaviour of onlyone species. An individual species can be absent due toreasons other than air pollution and even when present,only indicate one specific level of air pollution due to itsspecific SO2-sensitivity.

Correlation analysisSpearman rank order correlation coefficients were

calculated between average SO2 winter concentrationsover the last 5 years (1990-94) and IAP values. At the5 % significance level the correlation was consideredsignificant.

ResultsA total of 58 lichen species and 27 bryophytespecies were found at the 57 samplingstations used for evaluation.

Correlation with SO2 dataSpearman rank order correlation coeffi-

cients are listed in Table 4. With the excep-tion of IAP VI. 1 to VI.3, all IAP formulaetested gave a significant correlation withmean SO2 winter values over the last fiveyears. In spite of a substantial drop of emis-sions in the last decade, SO2 pollution stillseems to be the main factor determining thepresence or absence of epiphytes in the studyarea. In case of IAP VI however some valuesgave a positive, though not significant corre-lation with SO2 pollution (Table 4).

In general the use of a quantitative esti-mate of cover and/or abundance (IAP V upto IAP IX) did not increase the correlation.

Within formulae, a distinction can bemade on the basis of the species used. Fromresults of IAP I, it is clear that a selection ofbio-indicators (BI) gives better results thanall lichens. In turn, all lichens give a bettercorrelation than all epiphytes:

all epiphytes < all lichens< selection of bio-indicators

This trend is confirmed by most of the otherIAPs: correlation is almost always higher

TABL

E 4.

Sp

earm

an ra

nk

ord

er co

rrel

atio

n be

twee

n 91

lA

P-for

niulae

and

mea

n SO

,-con

cent

ratio

ns in

w

inte

r (O

ctobe

rMa

rch) o

ver

the 5

year

s 19

90

1994

(5

7 sam

ples

)

Aut

hor(s

)IA

P-fo

rmul

a+.

Spec

ies

use

d M

ean

1 tr

ee

10 tr

ees

Sam

plin

g gr

id

to o o

van

D

obbe

n (1

990)

, van

H

erk

(199

3)H

offm

ann

(1

993)

Hof

fman

n

(199

3)D

e Sl

oove

r &

Le

blan

c (1

968)

Her

ben&

Lis

ka

(198

6)

Her

ben

&

Li

ska (1

986)

Lebl

anc &

D

e Sl

oove

r (1

970)

Gee

bele

n (1

995)

Brit

t (1

987)

, N

imis

et a

l. (1

990)

Tra

ss (1

968,

1973

)

Ruo

ss et a

l. (1

991)

Haw

ksw

orth

&

R

ose (1

970)

Mea

n

IAPI

.l =

Xlic

hens

A

ll lic

hens

-0-

537

-0-

51*

-0-

58*

-0-

52*

IAP

1.2

=Se

piph

ytes

A

ll ep

iphy

tes

-0-

500

-0-

46*

-0-

52*

-0-

52*

IAP

1.3

=IB

I 17

lic

hens

-0-

583

-0-

59*

-0-

60*

-0-

56*

IAPI

I.l

=nxIQ

,

All

liche

ns

-0-

547

-0-

53*

-0-

58*

-0-

53*

IAPI

I.2

=n

xIQ

o

All

epip

hyte

s -0-

497

-0-

46*

-0-

52*

-0-

51*

IAPI

I.3

=n

xZ

QB

I 17

lic

hens

-0-

603

-0-

61*

-0-

62*

-0-

58*

IAPI

I.4

=

x

PH

17 lic

hens

-0-

623

-0-

65*

-0-

63*

-0-

59*

IAPI

II.l

=1

Q,

All

liche

ns

-0-

547

-0-

52*

-0-

59*

-0-

53*

IAP

III.2

=X

Q a

A

ll ep

iphy

tes

-0-

507

-0-

47*

-0-

53*

-0-

52*

IAP

III.3

=

Q B

I 17

lic

hens

-0-

603

-0-

61*

-0-

62*

-0-

58*

IAPI

II.4

=

P

H

17 lic

hens

-0-

650

-0-

68*

-0-

65*

-0-

62*

IAP

IV. 1

=l/

xX

Q,

All

liche

ns

-0-

460

-0-

44*

-0-

46*

-0-

48*

IAPI

V.2

=

l/

xIQ

u

All

epip

hyte

s -0-

550

-0-

54*

-0-

56*

-0-

55*

IAP

IV.3

=

l/x

ZQ

BI

17 lic

hens

-0-

643

-0-

66*

-0-

67*

-0-

60*

IAP

IV.4

=

1/M

XL

P B

All

liche

ns

-0-

467

-0-

42*

-0-

48*

-0-

50*

IAPI

V.5

=

l/

xZ

PB

17 lic

hens

-0-

447

-0-

36*

-0-

46*

-0-

52*

IAPI

V.6

=

1/M

XT

W

All

liche

ns

+0-

497

+0-

51*

+0-

45*

+0-

53*

IAP

IV.7

=

l/

xZ

Tw

17 lic

hens

+

0-50

0 +

0-42

*

+0-

54*

+0-

54*

IAPI

V.8

=

l/

xZ

PH

All

liche

ns

-0-

587

-0-

59*

-0-

57*

-0-

60*

IAPI

V.9

=

l/n

xIP

H

17 lic

hens

-0-

647

-0-

67*

-0-

64*

-0-

63*

IAP

V.I

= 1/1

0 x

EQ

,*/

All

liche

ns

-0-

530

-0-

54*

-0-

58*

-0-

47*

IAPV

.2

=l/

10

xZ

Qux

/ A

ll ep

iphy

tes

-0-

520

-0-

51*

-0-

58*

-0-

47*

IAPV

.3

=l/

10

xIQ

BIx

/ 17

lic

hens

-0-

543

-0-

57*

-0-

60*

-0-

46*

IAPV

.4

=l/

10

xIP

Hx

/ 17

lic

hens

-0-

633

-0-

65*

-0-

67*

-0-

58*

IAPV

I.l

=l/

nxIQ

,x/

All

liche

ns

-0-

18ns

- 0-

28ns

+

0-20

nsIA

P V

I.2

=\ln

x I

Q a xf

All

epip

hyte

s - 0-

02ns

- 0-

27ns

+

0-24

nsIA

PVI.

3 =

l/

xX

QB

,x

/ 17

lic

hens

-0-

13ns

-0-

15ns

+

0-12

nsIA

PVI.

4 =

l/

x!P

Hx

/ 17

lic

hens

-0-

580

-0-

60*

-0-

60*

-0-

54*

IAPV

II.l

=/

All

liche

ns

-0-

45*

IAPV

II.2

=

/

17 lic

hens

-0-

39*

IAPV

III.

l =

(P

H x cover)/

total

cover

A

ll lic

hens

-0-

51*

IAP

VII

I.2

=1

(PH x co

ver

)/tota

l co

ver

17

lic

hens

- 0-

55*

IAP

IX.l

=to

tal

cover

x )/1

00

All

liche

ns

- 0-

48*

IAP

IX.2

= to

tal

cover

x

)/l00

17

lic

hens

- 0-

50*

IAP

X

=1

all s

peci

es in

H

&R

ta

bles

H

&R

sp

ecie

s - 0-

59*

IAPI

-IA

PV

-0-

551

-0-

540

-0-

571

-0-

541

I 5' o' 5' 3 n s- O a s s to Ul*

= sig

nific

ant

at th

e 5%

-leve

l; ns

= not

signi

fican

t at

th

e 5%

-leve

l.^Fo

r expl

anat

ion

se

e M

ater

ials

and

Met

hods

.

256 THE LICHENOLOGIST Vol. 33

when the formula is based on a selection ofbio-indicator species (BI), and also whenpoleophoby indices of Barkman (1958) andHoffmann (1993) and the toxitoleranceindex of Wirth (1991) are used (IAP IV).

All IAP-formulae using poleophoby in-dices can be ranked in order of increasingcorrelation. The use of the poleophobyindex PH, based on study area specific data,consistently gave the best results:

Qa-values (all spp.)

2001 Evaluation of bio-indication methodsGeebelen & Hoffmann 257acidifying SO2 air pollution in a certain area.Weighting criteria based on a priori and localobservations (selection of locally distin-guished bio-indicator species, poleophobyindex based on local distribution patterns)increase the correlation considerably.

The use of a quantitative estimate of coverand/or abundance (frequency) does notresult in a higher correlation with SO2-concentrations. Liebendorfer et al. (1988)found a high regression coefficient betweenspecies frequency and SO2 and dust precipi-tation, while Herzig et al. (1985) and Herzig& Urech (1991) found the IAP summing allfrequencies to provide the best correlationwith long-term averages of air quality data.In Germany a standard guideline for airpollution monitoring uses species frequency(VDI 1995) which correlates well withseveral air pollutants (Kirschbaum &Windisch 1995; Kirschbaum 1995). Herben& Liska (1986) and Hoffmann (1993) onthe other hand found that implementation ofa quantitative factor is not necessary. Fromour results we conclude that abundance orcover estimates supply no extra informationabout the relation between corticolouslichens or epiphytes in general and SO2pollution. What might explain the discrep-ancy between the results obtained inSwitzerland (Liebendorfer et al. 1988;Herzig et al. 1985; Herzig & Urech 1991)and our own? Our research was carried outin an area with a strong urbanized character;even the stations considered as relativelyunpolluted, are still subject to considerableSO2 pollution or were so in recent history.This results in a severely impoverished epi-phyte flora, where most species are presentin low frequencies. The impoverishment ofCentral European corticolous vegetationseems to be less severe than in Flanders.This might explain why in our study area thepresence or absence of a species appears farmore indicative of the general air qualitythan its abundance or cover. Finally, unlikethe study areas in lowland Switzerland, usedby Herzig et al. (1985) and Herzig & Urech(1991), epiphytic vegetation in Flanders hasnever been very rich in species (see e.g.Kickx 1867; Duvigneaud 1942; Hoffmann

1993). This is due to less favourable climaticconditions and early human impact. More-over, there is a clear syntaxonomicaldifference between Flemish epiphyticvegetation (Hoffmann 1993) and lowlandSwiss epiphytic vegetation.

Among all the formulae tested, it is clearthat diose based on lichen data gave a highercorrelation than those using total number ofepiphytes (lichens+bryophytes). The bestresults were found when a selection of 17lichen species (BI) was used. These wereselected as eurytopic species (species with awide general ecological amplitude; e.g. wideclimatic and NH3-exposure tolerance), butwith strict constraints as far as acidifying airpollution is concerned (Hoffmann 1993). Itis important to note that this combination ofspecies was chosen using knowledge of therelation between their distribution patternsand the pollution pattern within part of thepresent study area.

This trend was confirmed when poleo-phoby indices were used i.e. when the QBIand PH indices were used based on only theselected bio-indicator species, a higher cor-relation was found than when using Q, andQa-values. Here also IAP values based onepiphytes (Qa-values) gave a lower corre-lation than those based on corticolouslichens (Qrvalues). Species might reactdifferently towards air pollution in otherregions that differ climatically or bio-geographically from the present study area.Therefore, it is recommended to use alllichen data in general instead of a selectionof bio-indicator species, which were selectedon the basis of their area-specific distributionas opposed to area-specific air pollutionpatterns.

From the IAP values obtained with thesecond formula of Herben & Liska (1986), itis also clear that application of the poleo-phoby index of Hoffmann (1993) gavebetter results than die toxitolerance indicesof Wirth (1991) and the poleophoby indexof Barkman (1958). This is again partiallyexplained by the fact diat the indices ofHoffmann are based on the relation betweendie distribution of the species and the esti-mated pattern of SO2-concentration, while

258 THE LICHENOLOGIST Vol. 33

Wirth (1991) and Barkman (1958) basedtheir indices on general experience. Further-more, the index of Hoffmann was based ondata from part of the present study area,while the indices of Barkman (1958) andWirth (1991) were based on experience inother (though neighbouring) areas.

Sampling 10 trees gave slightly better cor-relations than the use of a grid on the richesttree or sampling one tree. Compared to gridsampling, complete sampling of 1 and 10trees (up to 2 m high) has the disadvantageof introducing greater ecological variance.For example, species occurring at the base ofthe trunk can give an erroneous indication ofpresent air quality, owing to soil dust andmanure from cattle or urinating dogs (vanDobben 1987). Sampling all around thetrunk causes a supplementary ecological dif-ferentiation with strong influence on theabundance and frequency estimate. Only thealga Desmococcus olivaceus is generally foundall around the trunk of well-exposed trees,most other species are generally restricted tothe rain-exposed site of the trunk. Thisecological variation, not attributable to airpollution, is avoided when a sampling grid isused. Application of a sampling grid takesalso less time than a complete investigationof 10 trees. To determine the richest tree,a quick overview'of the ten trees is howeverrequired. One might therefore considerusing the most luxuriantly overgrowntree trunk instead of the species richestphorophyte.

In future SO2-monitoring within the studyarea, preference should be given to investi-gate the number of bio-indicator specieswithin a sampling grid. This method gives areasonably high correlation with all SO2-pollution data, is less time consuming, doesnot demand much lichenological knowledgeand is easy to calculate.

In other areas, however, these bio-indicator species should first be determined,as was done for lowland Switzerland byHerzig & Urech (1991). A possible alterna-tive is to include all lichens present. Werecommend sampling with a sampling gridon the tree carrying the highest number oflichen species.

We are grateful to the former Institute of Hygieneand Epidemiology (IHE) and the VlaamseMilieumaatschappij for providing the data on SO2deposition and to Dr M. Ciscato for critical reading ofthe manuscript.

R E F E R E N C E SAckermann, I. (1993) Flechtenkartierung im Raum

Leipzig. Flechten-monitoringein kommunalesKontrollinstrument: 21-34.

Aptroot, A. & Roos, R. (1993) Korstmos-oases inAmsterdam. De Levende Natuur 94: 130-134.

Barkman, J. J. (1958) Phytosociology and Ecology ofCryptogamic Epiphytes. Assen: Van Gorcum.

Barkman, J. J. (1963) De epifytenflora en -vegetatie vanMidden-Limburg (Belgie). Verhandeling der Konink-lijke Nederlandse Akademie van Wetenschappen, Afde-ling Natuurkunde, Tweede Reeks 54: 1-46.

Britt, H. (1987) Bioindikation mit Flechten am Beispielder Flechtenkartierung in Winterthur. Vierteljahr-schrift der Naturforschenden Gesellschaft in Zurich132: 123-134.

Caekebeke, G. (1986) Vergelijkende studie van de epi-fytische lichenen van de Denderstreek: 1962-1985.Dissertation, Universiteit Ghent.

De Bakker, A. J. (1984) Herkolonisatie van Zuidwest-Nederland door epifytische lichenen bij dalendeSO2-concentraties. Buxbaumiella 16: 18-23.

Deruelle, S. (1978) Etude comparee de la sensibilite detrois methodes d'estimation de la pollution atmos-pherique, en utilisant les lichens comme indi-cateurs biologiques, dans la region de Mantes(Yvelines). Revue Bryologique et Lichenologique 44:429-441.

De Sloover, J. & Lambinon, J. (1965) Contribution al'etude des lichens corticoles du bassin de laDendre. Bulletin de la Societe Royale de Botanique deBelgique 98: 229-273.

De Sloover, J. & Leblanc, F. (1968) Mapping ofatmospheric pollution on the basis of lichen sensi-tivity. In Proceedings of the Symposium on RecentAdvances in Tropical Ecology (R. Misra & B. Gopal,eds): 42-56. Varanasi: The International Societyfor Tropical Ecology.

De Wit, T. (1976) Epiphytic lichens and air pollution inThe Netherlands. Leersum: RIN-Verhandeling 8.

Diederich, P. & Serusiaux, E. (2000) The lichens andlichenicolous fungi of Belgium and Luxembourg: anannotated checklist. Luxembourg: Musee Nationald'Histoire Naturelle.

Duvigneaud, P. (1942) Les associations epiphytiquesde la Belgique. Bulletin de la Societe Royale deBotanique de Belgique 74: 32-53.

Fenton, A. F. (1960) Lichens as indicators of atmos-pheric pollution. Irish Naturalists' Journal 13:153-159.

Geebelen, W. (1995) Onderzoek naar de relatie tussenreele luchtvervuilingsmetingen en het voorkomen vanepifyten in Vlaanderen. Dissertation, UniversiteitGhent.

Gilbert, O. L. (1965) Lichens as indicators of airpollution in the Tyne Valley. In Ecology and the

2001 Evaluation of bio-indication methodsGeebelen & Hoffmann 259Industrial Society (G. T. Goodman, R. W. Edwards& J. M. Lambert, eds): 35-47. Oxford: UniversityPress.

Griffith, J. L. (1966) Some aspects of the effect of atmos-pheric pollution on the lichen flora to the west ofConsett, Co. Durham. M.Sc. Thesis, University ofDurham.

Hawksworth, D. L. & Rose, F. (1970) Qualitative scalefor estimating sulphur dioxide air pollution inEngland and Wales using lichens. Nature 227:145-148.

Haugsja, P. K. (1930) Uber den Einfluss der StadtOslo auf die Flechtenvegetation der Baume. NytMagazin for Naturvidenskabeme 68: 1-116.

Herben, T. & Liska, J. (1986) A simulation study onthe effect of flora composition, study design andindex choice on the predictive power of lichenbioindication. Lichenologist 18: 349-362.

Herzig, R., Liebendorfer, L. & Urech, M. (1985)Flechten als biologische Indikatoren der Luftver-schmutzung in der Schweiz (NFP14): Methodenent-wicklung in der Region Biel-Seeland. Systematisch-Geobotanisches Institut der Universitat Bern.

Herzig, R. & Urech, M. (1991) Flechten als Bio-indikatoren. Integriertes biologisches Messsystemder Luftverschmutzung fur das SchweizerMittelland. Bibliotheca Lichenologica 43: 1-283

Hoffmann, M. (1993) Verspreiding, ecologie en fyto-sociologie van epifyten en epifytengemeenschappen inOost- en West-Vlaanderen. PhD thesis, UniversiteitGhent.

Hill, M. O. (1979) TWINSPAN-A Fortran program forarranging multivariate data in an ordered two-waytable by classification of the individuals and attributes.New York: Cornell University.

Iserentant, R. & Margot, J. (1963) Onderzoek naarde luchtbezoedeling in de Kanaalzone Gent-Terneuzen; deel 2. Biologische enquete.

Institute of Hygiene and Epidemiology (1983-1993a)Belgisch automatisch meetnet voor luchtveron-treiniging. Brussels: Institute of Hygiene andEpidemiology.

Institute of Hygiene and Epidemiology (1983-1993*)Studie van de luchtkwaliteit in Belgie, Zwavel-rookmeetnet, jaarrapport. Brussels: Institute ofHygiene and Epidemiology.

Institute of Hygiene and Epidemiology (1990-1992)Metingen van de luchtkwaliteit in natuurgebieden.Jaarrapport. Brussels: Institute of Hygiene andEpidemiology.

Jones, E. W. (1952) Some observations on the lichenflora of tree boles, with special reference to theeffect of smoke. Revue Bryologique et Lichenologique21: 96-115.

Kershaw, K. A. (1985) Physiological Ecology of Lichens.Cambridge: Cambridge University Press.

Kickx, J. (1867) Flore cryptogamique des Flandres. Ghent:Librairie Generale de H. Hoste.

Kirschbaum, U. (1995) Immissionbezogene Flech-tenkartierung von Hessen nach der neuenRichtlinie VDI 3799 Blatt 1. StaubReinhaltungder Luft 55: 305-309.

Kirschbaum, U. & Windisch, U. (1995) Beurteilungder lufthygienischen Situation Hessens mittels epi-phytischer Flechten. Umweltplanung, Arbeits- undUmweltschutz. Schriftenreihe der Hessischen Landes-anstaltfiir Umwelt, Heft Nr. 171.

Leblanc, F. & De Sloover, J. (1970) Relation betweenindustrialization and the distribution and growthof epiphytic lichens and mosses in Montreal.Canadian Journal of Botany 48: 1485-1496.

Liebendorfer, L., Herzig, R., Urech, M. & Ammann, K.(1988) Evaluation und Kalibrierung der SchweizerFlechten-Indikationsmethode mit wichtigen Luft-schadstoffen. Staub-Reinhaltung der Luft 48: 133-138.

Londo, G. (1984) The decimal scale for releves ofpermanent quadrats. In Sampling methods and taxonanalysis in vegetation science, Handbook of VegetationScience, Vol. 4 (R. Knapp, ed.): 45-49. The Hague:Junk Publishers.

Nimis, P. L., Castello, M. & Perotti, M. (1990) Lichensas biomonitors of sulphur dioxide pollution in LaSpezia (Northern Italy). Lichenologist 22: 333-344.

Nylander, W. (1886) Les lichens du Jardin duLuxembourg. Bulletin de la Societe de Botanique deFrance 13: 164-372.

Oksanen, J., Laara, E. & Zobel, K. (1991) Statisticalanalysis of bio-indicator value of epiphytic lichens.Lichenologist 23: 167-180.

Quanten, E. (1986) Vergelijkende studie van de epifyten-flora in Midden-Limburg: 1960-1985. Dissertation,Universiteit Ghent.

Ruoss, E., Vonarburg, C. & Joller, T. (1991) Moglich-keiten und Grenzen der Flechtenbioindikation beider Bewertung der Umweltsituation in der Zen-tralschweiz. Bioindikation ein wirksames Instrumentder Umwelltkontrolle, VDIBerichteNi. 901: 81-101.

Ruoss, E. (1992) Flechten im Kanton Luzern. Unter-suchungen zur Bioindikation und Floristik, sowiezum Immissionsokologie voralpiner Hochmoore.Verb'ffentlichungen des Natur-Museums Luzern 3: 1-98.

Sergio, C. & Bento-Pereira, F. (1981) Liquenes ebriofitos como bioindicadores da poluicao atmos-ferica I. Utilizacao de uma escala qualitativa paraLisboa. Boletim da Sociedade Broteriana, Serie 2 54:291-303.

Sernander, G. (1926) Stockholms natur. Uppsala:Almqvist & Wiksells.

Showman, R. E. (1975) Lichens as indicators of airquality around a coal fired power generating plant.Bryologist 78: 1-6.

Showman, R. E. (1988) Mapping air quality withlichens, the North American experience. Bibli-otheca Lichenologica 30: 67-90.

Skorepa, A. C. & Vitt, D. H. (1976) A quantitative studyof epiphytic lichen vegetation in relation to SO2 pollu-tion in western Alberta. Edmonton: EnvironmentCanada [Information Report NOR-X-161].

Stalfelt, M. G. (1939) Der Gasaustausch der Flechten.Planta 29: 11-31.

Tanghe, M. & Richel, T. (1996) Premiere approche dela flore macrolichenique corticole de la regionde Bruxelles-capitale en situation de depollution auSO2. Belgian Journal of Botany 129: 38-46.

260 THE LICHENOLOGIST Vol. 33

Trass, H. (1968) Samblikud-ohu saastatuse indika-atorid. Eesti Loodus 11: 80-83.

Trass, H. (1973) Lichen sensitivity to the air pollutionindex of poleotolerance (I.P.). Folia CryptogamicaEstonica 3: 19-22.

Van der Gucht, K. & Hoffmann, M. (1990) The impactof air pollution on the occurrence of corticolousand saxicolous lichens in the industrial area Northof Ghent (Belgium). Memoires de la Societe Roy alede Botanique de Belgique 12: 111-126.

van Dobben, H. F. (1987) Effekten van ammoniakop epifytische korstmossen. In Acute en chronischeeffecten van NH3 (en NH^J op levende organismen(A. W. Boxman & J. F. M. Geelen, eds): 52-60. Nijmegen: Laboratorium voor AquatischeOecologie.

van Dobben, H. F. (1990) Effekten van luchtveront-reiniging op korstmossen, resultaten van meer-jarige studies. De Levende Natuur 91: 179-183.

Van Haluwyn, C. & Lerond, M. (1988) Licheno-sociologie et qualite de l'air: Protocole operatoire

et limites. Cryptogamie, Bryologie et Lichenologie 9:313-336.

van Herk, C. M. (1993) Korstmossen en zure depositiein Drenthe en Friesland. Hoofdrapport. Drenthe:Dienstwater en Milieuhygiene.

VDI-Richtlinie 3799, Blatt 1 (1995) Ermittlung undBeurteilung phytotoxischer Wirkungen vonImmissionen mit Flechten: Flechtenkartierung.

Vlaamse Milieumaatschappij (1993) Metingen vande luchtkwaliuit in natuurgebieden. Jaarrapport.Brussels: V.M.M.

Vlaamse Milieumaatschappij (1993-1994) Luchtveron-treiniging in het Vlaams Gewest, jaarverslag immis-sienet. Brussels: V.M.M.

Walther, H. & Lieth, H. (1967) Klimadiagramm-Weltatlas. Jena: Fischer.

Wirth, V. (1988) Phytosociological approaches to airpollution monitoring with lichens. BibliothecaLichenologica 30: 91-108.

Wirth, V. (1991) Zeigerwerte von Flechten. ScriptaGeobotanica 18: 215-237.

Accepted for publication 29 December 2000