I wanted to assess significant correlation between spatial variation in assemblages of species and higher taxa by using the RELATE routine in PRIMER; in which, Spearman’s correlation coefficients between the two similarity matrices would be determined. Matrices were constructed by both Bray–Curtis and Jaccard similarity measures of the presence/absence data of species, genera, families, orders, classes, and phyla. The results of spearman correlation for comparing matrices of p/a data of species and higher taxa by both Bray–Curtis and Jaccard similarity measures showed the same amounts; for example spearman rank for species vs. genus for both Bray–Curtis and Jaccard matrices is 0.975.
It is only appropriate to use the Jaccard index when dealing with P/A data but in my experience there is very little change in turnover patterns when comparing the two. There might be a larger differences when dealing with smaller communities. Bray-curtis should be used when your data take into account abundances.
Agree with Sean, use just one... Maybe BC is used more often and is part of mosty multivariate analyitical packages.
Hi all,
I was now finishing a reviewer report for a paper that compares both indexes. Sometimes, depending on your scientific question this is perfectly OK. Arguably, since BC takes abundances into account, it should be more sensitive to the processes that may underly the patterns you observe. I think it is always healthy to compare both indexes. However, if you are obtaining consistently the same results for the two indexes, I would go for the simple one when wrapping up a paper. You can always say the typical sentence. "We also used BC index (results not shown)... and come to the same conclusions".
You can report one, and simply note that the other gets similar results.
At large spatial scales (e.g., comparing regions) biodiversity is assessed usually by presence/absence and Jaccard index is most commonly reported. Although very similar, Sorensen or BC is used more often with at smaller scales such as comparing assemblages or 'communities' within a region; at this scale it is possible to quantify abundance.
When I analyse dataset with presence/absence data I prefer to use Jaccard index. As said by others, Bray-Curtis index fits better with abundance data.
Thank you all. Your comments helped me a lot. Since I used presence/absence data for both indices and according to comments I think it is better to use the suggested sentence "We also used BC index ... and come to the same conclusions"
Thanks again.
Jaccard es cualitatvie and BC es cuantitative.
For that reason if you have present/absence use Jaccard and for abundance use BC
I will recommend using alfa, beta, gamma methodology. Check the work of Gonzalo Halffter and Claudia Moreno from INECOL, Mexico.
Yes I think that is better to separate the diversity in their threee levels as Rony suggested.
In order to compare the different composition of the datasets, I will suggest to use beta diversity.
Your can read:
Koleff, P., Gaston, K. J., & Lennon, J. J. (2003). Measuring beta diversity for presence-absence data. Journal of Animal Ecology, 72(3), 367–382. doi:10.1046/j.1365-2656.2003.00710.x
If you want to reflect ecological distances, Kulcynsky and Relativized Manhattan as well as Bray-Curtis, are robust linear (proportional) mesures of ecological distance. See Faith et al. 1987. Vegetatio 69: 57-68.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA