~~~

 

Part 1. Select clusters of diatoms in the GoA between 2020-22

  

dia.bc.dist <- vegdist(otugen, method = "bray")
clusfd.KM.cascade <- cascadeKM(dia.bc.dist, inf.gr = 2, sup.gr = 10, iter = 100, criterion = "ssi")

 

This command can give different results, here it suggests 4 clusters (mostly recommends between 4, 9, and 10 clusters)

 

Then NMDS ordination shows which dates are clustered together

 

 

Based on the clusters, creating co-occurrence and mutual exclusion network of diatoms and: (i)top 30 18s genus, (ii)top 30 16s genus

 

Part 2. Create co-occurence & mutual exclusion network for each cluster

 

cluster 1: Early and mid mixing

 

18S

 

16S

 

 

cluster 2: Spring bloom

 

18S

 

 

16S

 

 

cluster 3: summer, autumn

 

18S

 

 

16S

 

 

cluster 4: bloom termination, summer, autumn

 

18S

 

 

16S