####環境統計学ぷらす第6回 #初版(20131205) data6.1<- read.csv("data6.1.csv",T) pairs(data6.1[,-1],panel=panel.smooth) pairs(data6.1[,10:12],panel=panel.smooth,cex=2) library(vegan) #PCA pca6.1<- rda(data6.1[,-1],scale=T) #prcomp(data6.1[,-1],scale=T) prcompによる解析も summary(pca6.1) biplot(pca6.1) data(varespec) varespec #CA ca6.1<- cca(varespec) summary(ca6.1) plot(ca6.1) #DCA dca6.1<- decorana(varespec) summary(dca6.1) plot(dca6.1) #RDA data(dune) data(dune.env) rda6.1<- rda(decostand(dune,"hell")~A1+Manure,dune.env) summary(rda6.1) plot(rda6.1) anova(rda6.1,by="margin") RsquareAdj(rda6.1) rda6.1.1<- rda(decostand(dune,"hell")~A1+Condition(Manure),dune.env) varpart(decostand(dune,"hell"),~A1,~Manure,data=dune.env) #CCA cca6.1<- cca(dune~A1+Manure,dune.env) summary(cca6.1) plot(cca6.1) anova(cca6.1,by="margin") #cluster解析 dis.bray<- vegdist(dune,"bray") dis.sorensen<- dist.binary(dune,method=5) dis.chao<- vegdist(dune,"chao") clus6.1b<- hclust(dis.bray,"average") clus6.1s<- hclust(dis.sorensen,"average") clus6.1c<- hclust(dis.chao,"average") plot(clus6.1b) plot(clus6.1s) plot(clus6.1c) #NMDS nmds6.1<- metaMDS(dis.chao) #nmds6.1<- metaMDS(dune,"chao") でもOK nmds6.1 plot(nmds6.1,type="t") ordihull(nmds6.1,dune.env$Management) #Mantel検定 dis.env<- dist(dune.env$A1) #一変量でやっているが、多変量で非類似度計算してもOK mantel(dis.chao,dis.env) #ANOSIM anosim(dis.chao,dune.env$Management) #PerMANOVA adonis(dis.chao~dune.env$Management)