~~~

Ordination

 

Exploring NMDS ordination based on Bray-Curtis dissimilarity. Testing differences between log transform, and double cutoff of the data. Also looking on the difference between the total data (not log-transform), and only surface data (not log-transform).

Note: Because the data shows concentration and not counts, maybe log-transform is acceptable. Note 2: The clusters of the un-transformed data that we discussed was not showing 70% similarity, but lower. Here this data is updated, and based on 65% similarity. It added another cluster of the mixing- twoards bloom period.

 

Data log-transform only

Ordination based on similarty of 70% or higher.

 

 

 

 

Data log-transform and double cutoff

Ordination based on similarty of 70% or higher.

 

 

 

 

Data double cutoff, not log-transform

Ordination based on similarity higher than 65%.

 

 

 

 

Data of surface samples, double cutoff, not log-transform

Ordination based on similarity higher than 65%.