Treść książki
Przejdź do opcji czytnikaPrzejdź do nawigacjiPrzejdź do informacjiPrzejdź do stopki
16
20Methods
Thefollowingmethodsofanalysiswereapplied:C&RT–Classificationand
RegressionTree(Breimanetal01993;Ripley1996;Loh,Shih1997),PCA(Prin-
cipalComponentsAnalysis),RDA(RedundancyAnalysis),CCA(Canonical
CorrespondenceAnalysis)0Thecalculationswereperformedintheprograms
STATISTICA800-12,CANOCO405andPAST0
Detrendedcorrespondenceanalysis(DCA)wasperformedtoexzamine
thegradientlengthofvegetationdiversity0ThelengthofgradientforAxis
Iof304SDgivesgreatflexibilitytochoosethemethodofdirectordination(ter
Braak,Šmilauer2002)0
Toshowthediversityofenvironmentalconditionsinthestudiedlakes
PCAwasusedinthePAST2012program04650environmentalsampleschar-
acterisedby26featureswereanalysed0Thedatawerestandardiseddueto
differentmeasuresoffeatures0Correlationsbetweenenvironmentalfeatures
werecalculatedbynon-parametricSpearman’srankcorrelationcoefficient
(Spearman’s(rho)correlation;Siegel,Castellan1988;STATISTICA12)0Statisti-
callysignificantcorrelationsareatp<000010
Theeffectofenvironmentalconditionsonspeciesrichness,emergentand
submergedplants,vascularplants,stonewortsandmosses(numberofspecies
andbiomass)wasdeterminedusingRDAonthebasisofthecalculationsmade
inCANOCO405(Hastieetal02001;terBraak,Šmilauer2002)0Beforetheanalyses
focusedonscalingoninter-speciescorrelations,sampleswerecentredand
standardised,specieswerecentred,butthedatawerenottransformed0Se-
lectionofenvironmentalvariableswasautomatic0Thesignificanceofdiffer-
enceswastestedbyMonteCarlopermutationtests(terBraak,Šmilauer2002)0
Therelationbetweenspeciesdiversity,plantbiomass,numberofspeciesand
waterandsedimentfeatureswascalculatedbynon-parametricSpearman’s
rankcorrelationcoefficient(Siegel,Castellan1988;STATISTICA12)0Statistically
significantcorrelationsareatp<000010
Theevaluationoftheimportanceofenvironmentalvariablesinthediffer-
entiationofthedistinguishedcommunitiesofmacrophyteswasperformed
bydiscriminantanalysis0TheCanonicalCorrespondenceAnalysis(CCAfrom
theCANOCO405program)wasused0Speciesdatawastookintoaccountallthe
plantsamples(10692)classifiedas12communities,andenvironmentaldata
included21featuresofwaterandsedimentsamplescollectedfrom775depth
zones0IntheanalysesHill’sscalingtypewasused,aswellasautomaticselec-
tionofenvironmentalvariablesandtheMonteCarlopermutationtest0
Theanalysisofrelationshipsbetweencommunitiesandenvironmental
conditionsinlakeswasperformedbymeansoftheClassificationandRe-
gressionTree(C&RT)withGiniimpurity(I
G(f))intheSTATISTICA800software,
followingrecommendationsgivenbyBell(1999);Austin(2007)andOlden
