Dia om assemblages dis ibu ion in Ca alan i e s, NE Spain, in ela ion o chemical and
physiog aphical ac o s
Manel Lei a1* and Se gi Saba e 2
1Cen e o he En i onmen & Depa men o Geog aphy Uni e si y o Dublin, T ini y
College, Dublin 2, I eland; Tel: +353-1-6083876; Fax: +353-1-6713397; E-mail: [email p o ec ed]
2Depa amen de Ciències Ambien als, Facul a de Ciències, Campus de Mon ili i, Uni e si a
de Gi ona, E-17071 Gi ona, Spain; Tel: +34-972 418 700: Fax: +34-972 418 150; E-mail:
[email p o ec ed]
Keywo ds: dia oms, epili hon, wa e quali y, mul i a ia e me hods, communi y analyses.
Abs ac
Dis ibu ion pa e ns o epili hic dia om assemblages in s eams in no heas e n Spain and hei
ela ion o di e en en i onmen al g adien s a e p esen ed. Thi y- i e si es we e sampled
co e ing a wide ange o lu ial ypologies. G adien analysis was used o analyse he
communi y s uc u e and he majo ecological g adien s unde lying a ia ion in species
composi ion. Two majo g adien s we e e iden ; he i s was a complex g adien om
oligo ophic, p is ine, as - lowing highland i e s o mainly eu ophic i e s o low ele a ion;
and he second ela ed o al i udinal and seasonal a ia ion o empe a u e. Two se s o ac o s
we e e iden ; one was associa ed wi h wa e chemis y, and he o he o seasonal and
physiog aphical a ia ion. Va iance pa i ioning allowed he sepa a ion o he e ec s o he
di e en se s o en i onmen al pa ame e s. The con ibu ion o physiog aphy and wa e
chemis y o dia om dis ibu ion was mo e unclea when he le el o dis u bance was
in e media e. Conside ing he ela i ely high p opo ion o a ia ion explained by
physiog aphical a iables alone, we sugges ha when using dia oms o e alua e wa e quali y,
eco egional cha ac e is ics o i e s e ches should also be conside ed.
1. In oduc ion
* Co esponding au ho : Cen e o he En i onmen & Depa men o Geog aphy Uni e si y o Dublin, T ini y
College, Dublin 2, I eland; Tel: +353-1-6083876; Fax: +353-1- 6713397; E-mail: lei am@ cd.ie
1
Chemical analyses o wa e p o ide a good indica ion o he chemical quali y o he aqua ic
sys ems, bu do no in eg a e ecological ac o s such as al e ed ipa ian ege a ion o al e ed
low egime and he e o e, do no necessa ily e lec he ecological s a e o he sys em (Ka e
al. 2000). Biological assessmen is a use ul al e na i e o assessing he ecological quali y o
aqua ic ecosys ems since biological communi ies in eg a e he en i onmen al e ec s o wa e
chemis y, in addi ion o he physical and geomo phological cha ac e is ics o i e s and lakes
(S e enson and Pan 1999).
Dia oms a e a siliceous class o algae epu ed o being e y sensi i e o chemical condi ions.
They usually accoun o he highes numbe o species among he p ima y p oduce s in aqua ic
sys ems (Pan e al. 1999). Consequen ly, hey a e equen ly used as biological indica o s o
wa e quali y and a numbe o me hods o moni o ing Eu opean i e s based on dia oms ha e
been al eady p oposed (Kelly e al. 1998, P ygiel e al. 2002). Dia om analysis ha e he e o e
been conside ed an impo an con ibu ion o he Eu opean Wa e F amewo k Di ec i e
(2000/60/EC), which aims o achie e a “good s a us” o all wa e s in he EU be o e 2015.
Despi e hei po en ial, he use o dia oms as biological indica o s o wa e quali y o lu ial
sys ems may be hampe ed because hei dis ibu ional pa e ns espond o a mul i ude o
di e en ac o s, anging om biogeog aphical o biogeochemical and human in luenced
(Po apo a and Cha les 2002). The challenge o de e mining which po ion o he dis ibu ion o
a gi en communi y is associa ed wi h human in luences and which is ela ed o he de ining
cha ac e is ics o a gi en eco egion needs o be esol ed (e.g. Hall e al. 1999). This knowledge
may be essen ial in he assessmen o biological quali y, which is based on he deg ee o
de ia ion om a e e ence biological popula ion. Biological in eg i y (Ka 1991) is measu ed
by compa ing a gi en si e wi h e e ence ecosys ems ha lie in simila geomo phological and
clima ic se ings bu a e no exposed o human impac . The sea ch o e e ence si es is o en
di icul in pa icula in egions o ex ensi e human ac i i ies, and his may hampe he
applica ion o his concep . Knowing in dep h he au oecological cha ac e is ics o species may
be ex emely help ul o make an objec i e sea ch o e e ence condi ions.
2
Mul i a ia e echniques ( e B aak and Ve donscho 1995) allow he elucida ion o ecological
ac o s, which explain mos o he a ia ion in dia om dis ibu ion. In pa icula , a iance
pa i ioning (Bo ca d e al. 1992) may help o sepa a e egional and gene al ac o s in luencing
dia om dis ibu ion. In his pape we use mul i a ia e echniques o elucida e he
physiog aphical, chemical and physical ac o s associa ed wi h dia om communi y s uc u e in
NE Spain.
2. Me hods
2.1. Dia om sampling, p epa a ion and analysis
This s udy was ca ied ou in a numbe o i e s loca ed in Ca alonia (NE Spain) (Fig. 1). These
si es we e selec ed o co e a wide ange o lu ial ypologies wi h di e en le els o human
dis u bance. Fi y-se en epili hic samples we e collec ed du ing win e (Feb ua y) and sp ing
(May) 2001. Dia oms we e p epa ed by acid oxida ion wi h concen a ed sulphu ic acid and
po assium dych oma e. Pe manen slides we e moun ed using Naph ax ( .i. 1.74). A leas 400
al es we e coun ed on each slide on andom ansec s wi h a ligh mic oscope using phase-
con as and Noma ski di e en ial in e e ence con as op ics a a magni ica ion o 1000x o
1250x. The iden i ica ion mainly ollowed K amme and Lange-Be alo (1986-1991).
2.2. Si e cha ac e iza ion
The mo phological and clima ic di e si y o NE Spain de e mine an impo an spa ial
he e ogenei y in he egion. The Py enees and nea by a eas ha e clima es ypical o high
moun ainous egions, wi h minimum empe a u es below 0ºC, annual ain all o abo e 1,000
mm and hea y snow all in he win e . Coas al a eas ha e mild, empe a e wea he , wi h
empe a u es ha ise and ain all ha dec eases as one mo es om no h o sou h. Inland, a
om he sea, he clima e is ypical o con inen al Medi e anean egions, wi h cold win e s and
e y ho summe s. The si es we e selec ed o co e a la ge g adien anging om p is ine o
pollu ed wi h di e en le els o human dis u bance, co e ing he majo ypes o
geomo phological and physiog aphical condi ions.
Chemical da a (conduc i i y, suspended ma e , dissol ed oxygen, pH, BOD5, NH s up 7(+ )-N,
Ca2+, PO
Com
b i n
–P, Mg2+, NO s up 7(- ) -N, NO s up 7(- )-N, K+, Na+, SO
Com
b i n
, Cl-) o he si es we e ob ained
3
om he Ca alan Wa e Agency (ACA) which moni o s hem ou inely. Wa e samples we e
analysed ollowing s anda d p ocedu es (APHA 1989). The ollowing physical desc ip o s o
he si es we e used o his s udy: wa e empe a u e, al i ude, slope, s eam o de , ca chmen
a ea, dis ance o sou ce, and bi u ca ion a io (i.e. he quan i a i e a io de e mined be ween
he numbe o s eam segmen s in one o de and he nex ). Pa s o hese da a we e de i ed
om maps and ield obse a ions, while o he s we e p o ided by he ACA.
2.3. Da a analysis
Dia om da a we e analysed wi h de ended co espondence analysis (DCA) (Hill and Gauch
1980) o de e mine he leng h o he g adien o he i s wo axes. DCA indica ed ha he
g adien leng h was g ea e han 2 s anda d de ia ion uni s (3.191); he e o e, he use o
unimodal o dina ion echniques would be app op ia e ( e B aak 1987). Consequen ly,
co espondence analysis (CA), an indi ec –g adien o dina ion, was used o de e mine he majo
pa e ns o a ia ion in species composi ion da a wi hou inco po a ing da a on en i onmen al
a iables. Cons ained o dina ion (canonical co espondence analysis, CCA) was used o ela e
dia om assemblage s uc u e o all p edic o en i onmen al a iables and o explo e he
ela ionships among and be ween species and he en i onmen ( e B aak and Ve donscho
1995). Acco ding o his p elimina y CCA, we iden i ied collinea a iables and selec ed a
subse o a iables based on inspec ion o a iance in la ion ac o s (VIF<20) ( e B aak and
Šmilaue 1998). S ep-wise o wa d selec ion and Mon e Ca lo pe mu a ion es we e used o
educe u he he en i onmen al a iables o hose co ela ed signi ican ly wi h he de i ed
axes ( e B aak and Šmilaue 1998), a a cu -o poin o P=0.05. P obabili ies o mul iple
compa isons we e co ec ed using he Bon e oni co ec ion. Pa ial CCA was used o sepa a e
and examine he ela i e impo ance o he species da a o se e al se s o explana o y a iables
on he dia om assemblages (Bo ca d e al. 1992). We we e in e es ed in sepa a ing he e ec s o
he wa e chemis y om all he o he a iables and hen es ing whe he hese wo di e en
g oups o a iables we e edundan o each o he , o whe he hey each explained unique
aspec s o species composi ion.
4
Only hose axa ha we e obse ed in mo e han 5% o he samples we e included in analyses
o axa abundances o minimize he in luence o a e axa. Taxa abundance was squa e oo
ans o med in all analyses o educe he e ec o highly a iable popula ion densi ies on
o dina ion sco es. En i onmen al da a (excep pH and pe cen age slope) we e loga i hmically
ans o med be o e analysis o educe skewed dis ibu ions. All o dina ions we e pe o med
using CANOCO e sion 4.1 ( e B aak and Šmilaue 1998).
3. Resul s
3.1. En i onmen al cha ac e is ics o si es
Ri e s o he s udy a ea a ied widely in wa e quali y and habi a cha ac e is ics. Conduc i i y
inc eased downs eam and he highes alues o suspended solids we e ound a he lowe mos
si es. Nu ien concen a ions also inc eased downs eam bu a ied mo e spa ially han did
o he majo a iables because he ex ensi e applica ion o chemical e ilize s in ag icul u e and
di e se sewage inpu s (Table 1 & 2). Some s udy si es we e loca ed in densely popula ed,
hea ily indus ialized o in ensi e ag icul u al a eas, hus ecei ing high inpu s o o ganic
ma e and indus ial sewage. O he si es we e less dis u bed, as - lowing and wi h a s ony o
s ony-g a el bo om, cha ac e is ic o moun ainous a eas. Finally, o he s ha e a lowland
cha ac e , being slow- lowing.
3.2. Species composi ion o dia om assemblages
A o al o 195 axa we e ound. Only 33 axa we e obse ed in mo e han 5% o he samples.
Ampho a pediculus (Kü zing) G unow was he mos abundan axon and occu ed in up o 78%
o all samples. The second mos abundan species we e Achnan hes minu issima Kü zing and
Na icula g ega ia Donkin. CA o dina ion esul s showed ha 15.2% o dia om assemblages
a iance was explained on axis 1, and a u he 13% on axis 2. Si es on he le side o he i s
CA axis (Fig. 2) lay in densely popula ed, indus ialized o ag icul u al a eas, ecei ing high
inpu s o o ganic ma e and indus ial discha ge. Dia om axa showing maximum abundance in
samples om si es si ua ed on he le side o he diag am included Ni zschia capi ella a
Hus ed , Ni zschia us ulum (Kü zing) G unow and Na icula subminuscula Manguin. Si es less
in luenced by human ac i i ies had highe alues on CA axis 1 and a e clus e ed on he igh
5

hand side o he diag am. Dia oms abundan in hese samples included Achnan hes minu issima
Kü zing, Achnan hes biasole iana G unow and Gomphonema pumilum (G unow) Reicha d &
Lange-Be alo . The second axis is mos p obably e lec ing seasonal di e ences in he dia om
communi ies, since he win e samples had highe alues, while hose aken du ing sp ing ime
had lowe alues on CA axis 2.
3.3. Impo an en i onmen al a iables
The 21 en i onmen al a iables al oge he explained 52.0% (2.067 s.d. uni s) o he o al
a ia ion wi hin he species da a. The i s wo axes explained a signi ican po ion o a iance
in he dia om axa da a (P<0.01). The la ge a iance in la ion ac o s (>20) o he explo a o y
CCA indica ed edundancy among dis ance o sou ce, sulpha es, po assium, sodium,
magnesium, chlo ides, calcium and ca chmen ’s a ea, and we e subsequen ly excluded om he
en i onmen al da a se .
CCA wi h o wa d selec ion o species da a indica ed ha conduc i i y, wa e empe a u e,
BOD5, al i ude, s eam o de and ni a e ni ogen we e he en i onmen al a iables ha
accoun ed o signi ican (P<0.05) po ions o he o al a iance in dia om species composi ion
(Table 3). Conduc i i y explained he la ges po ion (7.3%) o he o al uncons ained a iance.
CCA o dina ion plo s (Fig. 3a, b) p ima ily exp essed wo majo g adien s. The i s one
a anged si es om oligo ophic, p is ine, as - lowing highland i e s o p edominan ly
eu ophic i e s o low ele a ion. The second g adien is likely in eg a ing a seasonal and
al i udinal a ia ion o empe a u e. Cyclo ella meneghiniana Kü zing, Na icula subminuscula,
Ni zschia us ulum and Ni zschia capi ella a ha we e mainly obse ed a si es wi h high
conduc i i y and BOD5, usually a lowe al i udes. Achnan hes minu issima and Ni zschia
on icola G unow we e ins ead associa ed wi h mo e p is ine wa e s and high al i ude si es.
3.4. Rela i e impo ance o chemical s. physiog aphical ac o s
Pa ial CCA analyses showed ha 25.5% (0.526) o he o al a ia ion in he species da a could
be explained by bo h se s o a iables oge he . Wa e chemis y explained a conside ably high
p opo ion (12.9%, 0.266) o he o al a iabili y in species da a. The ac ion o he o al
a iance a ibu ed o physiog aphical a iables was 11.6% (0.239). Since he alue o he
6
in e sec ion (0.021, 1.0%) was much smalle han he ac ion o he a ia ion explained
independen ly by bo h se s o a iables, we can conclude ha he wo se s o a iables we e
p ima ily explaining unique aspec s o species composi ion.
The summa y o he o dina ion gi es an o e all measu emen o how much a ia ion can be
ela ed o bo h explana o y a iables. Bu no all species a e equally well explained by he same
se o en i onmen al a iables. The i o species can be used as a measu emen o ind ou
which species a e well ep esen ed and he pe cen age o a iance i by each se o explana o y
a iables. The pe cen age o a iance explained by wa e chemis y and he physiog aphical
ac o s wi h espec o he uncons ained a iance di e s among he di e en species (Table 4).
Na icula a omus (Kü zing) G unow a . pe mi is (Hus ed ) Lange-Be alo , Ni zschia
capi ella a, N. us ulum and Cyclo ella meneghiniana had hei dis ibu ions mos ly explained
by he wa e chemis y. Cocconeis pediculus Eh enbe g and C. placen ula Eh enbe g,
Gomphonema oli aceum (Ho nemann) B ébisson and Ni zschia inconspicua G unow we e
amongs he mos ela ed o he physiog aphical g oup o a iables. Finally, Melosi a a ians
Aga dh, Na icula subminuscula, Cymbella minu a Hilse ex Rabenho s and Ni zschia on icola
we e bo h explained by he physiog aphical and chemical a iables.
4. Discussion
App oaches o he use o dia om communi ies as biological indica o s di e ge om he di ec
applica ion o dia om indices, which ha e been de eloped and widely used in Eu opean wa e s
(Descy and Cos e 1991, Kelly and Whi on 1995) o mul ime ics o mul i a ia e app oaches
(e.g. IBI o index o biological in eg i y), which a e p e e ed in he Uni ed S a es (Hill e al.
2000, Ka 1991). I espec i e o he me hod used, unce ain ies s ill exis conce ning he
uniqueness o gene al applicabili y o a gi en diagnosis p o ided by he dia om communi ies.
This is especially he case when pollu ion o nu ien en ichmen is no d i ing communi y
composi ion. Leland and Po e (2000) obse ed ha incomple e p edic ion o phospho us
loading by dia om communi ies occu ed sys ema ically a lowe phospho us concen a ions.
Saba e (2000) showed ha dia om indices success ully indica ed he e ec o a ca as ophic
hea y me al spill on a i e sys em, bu ailed o eliably de ec he eco e y. Thus in cases
7
whe e dis u bance is no ex eme, eco egional cha ac e is ics (e.g. geochemical ea u es,
empe a u e, wa e low) and esponses o en i onmen al s esso s (e.g.: nu ien en ichmen ,
habi a al e a ion, oxic inpu s) may combine o p oduce a con using pic u e.
The use o mul i a ia e analyses in he p esen s udy showed he g adien om p is ine o
pollu ed i e s as being he mos impo an in s uc u ing i e ine dia om assemblages o i e
wa e s in NE Spain. A esponse o dia om communi ies o nu ien en ichmen appea ed o be
e iden , al hough he CCA showed ha he e ec o nu ien s and conduc i i y on species
composi ion was simila . Many s udies ha e shown ha dia om communi ies a y along
g adien s o ionic concen a ion (Saba e and Roca 1992). Mos o en he inc ease in ionic
con en is accompanied by nu ien en ichmen (Saba e and Saba e 1988, Leland and Po e
2000). The downs eam en ichmen wi h dissol ed sal s, and nu ien s in pa icula , is
pa icula ly clea in his s udy. The o e iding e ec o conduc i i y and nu ien s on he
composi ion o dia om communi ies has been ecognized ecen ly h ough he analysis o la ge
da a se s (Po apo a and Cha les 2003). In hese ci cums ances ca e mus be aken no o con use
he e ec o ionic con en wi h ha o nu ien en ichmen .
I is impo an o s ess ha dia om dis ibu ion was no only a ec ed by chemical
cha ac e is ics. CCA showed ha dia om communi y s uc u e could also be ela ed o
empe a u e, al i ude and s eam o de . Gene al desc ip o s summa izing physical habi a (e.g.:
s eam size and s eam o de ) ha e been associa ed wi h dia om dis ibu ion (Molloy 1992).
Physiog aphical a iables, especially al i ude and s eam o de , may p o ide in o ma ion on
how i e basins a e s uc u ed and, analogously, on how dis u bances occu ing a he basin
scale (human ac i i ies, changes in land use o in ege a ion) may a ec dia om communi ies.
A smalle spa ial scales i has been also shown ha subs a um ype and habi a he e ogenei y
a e ela ed o species composi ion and p e e en ial mo phological g ow h o ms o dia om
communi ies (Leland 1995). A he basin scale used in his s udy, al i udinal g adien may be
pa ly ela ed o he ionic en ichmen o i e wa e s. Ma gale (1960) sugges ed ha algal
assemblages (including dia oms) we e ela ed o he geochemical and nu ien composi ion o
wa e , as well as o he hyd aulics and geomo phology o he i e sys em. This p edic ion
8
ad anced he indings based on he use o mul i a ia e analysis, as well as he p edic ions o he
Ri e Con inuum Concep (Vanno e e al. 1980). Ou analyses showed ha pa e ns o dia om
dis ibu ion could be a ibu ed no only o wa e quali y a ia ion, bu also o physiog aphical
p ocesses. The dis ibu ion o some common and abundan axa in ou da a se is la gely
in luenced by physiog aphical pa ame e s. Th ee di e en g oups o axa can be iden i ied wi h
ega d o hei link o he chemical, physiog aphical o mixed (o e lapped) in luences (Table 4).
Al hough mos o hese axa ha e widesp ead dis ibu ion elsewhe e, p e ious s udies sugges ed
ha hei occu ence is ela ed o speci ic ype o en i onmen s. Fo example, Gomphonema
oli aceum has a dis ibu ion es ic ed o low-o de s eams wi h high alkalini y and low
nu ien inpu (Pa ick 1977), indica ing s ong physiog aphical in luence. Na icula a omus a .
pe mi is and Ni zschia capi ella a occupied highly pollu ed sec ions o i e s, and we e
he e o e amongs he mos in luenced by he wa e chemical cha ac e is ics o he i e s. Thei
occu ence was i espec i e o he s eam o de o loca ion (Lange-Be alo 1979). Melosi a
a ians was ound mainly in downs eam sec ions o he s udied wa e cou ses, occupying
nu ien -en iched, mine alised wa e s (Ma gale 1951). This species is he e o e an exp ession
o he o e lapping in luence be ween he chemical and he physiog aphical ac o s.
5. Conclusions
A p esen he in luence o small-scale ac o s on dia om communi y s uc u e is poo ly
unde s ood. This makes i di icul o es ablish any s anda d me hodology o assess wa e
quali y in a wide geog aphical a ea, pa icula ly in hose si es wi h in e media e dis u bance.
Se ing e e ence s anda ds o biological quali y will equi e he ca e ul de elopmen o da a
se s along majo en i onmen al g adien s unde lying species composi ion, in o de o
unde s and he p ocesses ha a e s uc u ing dia om assemblages. We ha e demons a ed ha a
conside able p opo ion o a ia ion in he dia om assemblages can be explained by
physiog aphical ac o s; i would he e o e be mo e eliable o de elop dia om-based
app oaches o wa e quali y es ima ion ha a e applied a he eco egional le el o in
homogeneous geog aphical a eas, hus a oiding unwan ed e ec s on da a se s.
9