results SIMPER dissimilarity for species with equal mean abundance

[sofia]

Dear all,

I am doing my master thesis on subtidal algae species in the north Atlantic. Herefore I have presence-absence data of different algae species. I have a balanced design, with factors depthlevel (levels a&b), year (3 levels) and transect (10 levels) and 2 replicates per deplthlevel per transect per year. A permanova showed significant results for e.g. the factor year.

So I continued with a SIMPER analysis for dissimilarity on a bray-curtis similarity index, no transformation, to find the species contributing most to the differences between years. (I tried both the programs primer-e and Past)

and now my question:

But now I find when for example comparing between the first 2 years, that the species C. baccata contributes to the dissimilarity between groups (2009 and 2010), however it has the same average abundance.

does someone know if it is possible that this species contributes to the dissimularity (for 8.21%), but has the same average abundance? And if so, how? And if not, does someone have an idea on the cause and on how to solve this problem?

the average/mean (term depending on the program used) abundance can be seen as the proportion of samples with this species present i guess?


thanks for the effort, and any comments will be of great help!


output primer-e 2006 version:

Group 2009 Group 2010
Species Av.Abund Av.Abun Av.Diss Diss/SD Contrib% Cum.%
S. polyschides 0,33 0,78 6,55 1,16 9,90 9,90
Laminaria spp 0,43 0,70 5,75 1,02 8,70 18,60
C. baccata 0,50 0,50 5,43 0,96 8,21 26,81
A. armata 0,43 0,23 4,93 0,88 7,46 etc.

[bugman]

its probably an artefact of using presence absence data. Can you try re-running the analysis using the Sorenson similarity index. Does this result in the same output?