The correlation between the environment variable and the species composition is determined by canonical correlation analysis (CCA). For each sample, except for the data on the species composition and abundance, you must have data on environmental factors (eg, humidity, acidity, etc.). You should also keep in mind that: 1) CCA requires the multivariate normal and homogeneity of variance assumption, 2) assumes a linear relationship between the canonical variates and each set of variables, and 3) requires a large sample size. As a result, for each environmental factor you get a contribution to the explanation of the data distribution and the level of statistical significance.

Detrended correspondence analysis (DCA) is a multivariate statistical technique to find the main factors or gradients in large, species-rich but usually sparse data matrices that typify ecological community data. Look here http://ordination.okstate.edu/DCA.htm  Conclusions about the gradients of changes in the environment are made on the basis of species composition data. There must be carefully preparation of data and elimination of outliers. Keep in mind, too, that there is no opportunity to test the statistical significance.

The difference between MDS, PCA and other method of ordination well-reviewed here https://www.researchgate.net/post/How_to_choose_ordination_method_such_as_PCA_CA_PCoA_and_NMDS

Each analysis method has its own peculiarities and use in source data requirements, as well as advantages and disadvantages.

Best wishes

The correlation between the environment variable and the species composition is determined by canonical correlation analysis (CCA). For each sample, except for the data on the species composition and abundance, you must have data on environmental factors (eg, humidity, acidity, etc.). You should also keep in mind that: 1) CCA requires the multivariate normal and homogeneity of variance assumption, 2) assumes a linear relationship between the canonical variates and each set of variables, and 3) requires a large sample size. As a result, for each environmental factor you get a contribution to the explanation of the data distribution and the level of statistical significance.

Detrended correspondence analysis (DCA) is a multivariate statistical technique to find the main factors or gradients in large, species-rich but usually sparse data matrices that typify ecological community data. Look here http://ordination.okstate.edu/DCA.htm  Conclusions about the gradients of changes in the environment are made on the basis of species composition data. There must be carefully preparation of data and elimination of outliers. Keep in mind, too, that there is no opportunity to test the statistical significance.

The difference between MDS, PCA and other method of ordination well-reviewed here https://www.researchgate.net/post/How_to_choose_ordination_method_such_as_PCA_CA_PCoA_and_NMDS

Each analysis method has its own peculiarities and use in source data requirements, as well as advantages and disadvantages.

Best wishes

Fadli,

I am not sure what statistical software you are planning to use but I found Primer-E to be very good for your type of hypothesis testing. The web site is:

http://www.primer-e.com/primer.htm

Good luck,Tom

Please check these attachments for said query.

I agree with what Denis wrote about CCA, I would add that CCA assumed an unimodal, bell shaped, response between environmental variables and species composition, which usually is the case when you have large enough environmental gradients so that you have both suitable and unsuitable environments for your species. If instead you have short environmental gradients, so that you can assume a linear response between environmental variables and species composition, Redundancy Analysis (RDA) is usually recomended. RDA is very similar to PCA (is a multiple linear regression) but it features explicit dependence of response variables (species composition in your case) from environmental variables, whilst PCA handles all variables together.

I think the most complete manual avout those methods id Legendre and Legendre's Numerical Ecology, and all those methods are implemented in the R package Vegan, which in my opinion is very well done and easy to use.

Hi Fadli,

Because you probably have lots of zeros in your data set, and not a Normal distribution of your data, I'll use non-paramteric analysis. You can try to use PRIMER software. First run some MDS (Multidimesional Scaling) to look at species composition by site or habitat, and then you can correlate your environmental dataset to species composition data using Bio-Env analysis. Feel free to contact me if you have more questions.

I also favor NMDS, running both your environmental data and community composition in joint plots. You could also use cluster analysis to group your sites by your desired environmental gradient and then run indicator species analysis to figure out which species best represent your environmental groups.

I guessed this type of question was discussed here long before as well. If I remembered correctly, Jari Oksanen and Gavin Simpson were also a bit reluctant by their vegan package first by testing the axis length through DCA. If the value of axis length be 2.5 or longer (Palmer, Okostate,edu) better to use CCA otherwise stick to RDA. No PCA or CA will answer your question. They are simply for species and sample scores.

I used (Multivariate) Fuzzy Set Ordination in R for similar problems. Please check my articles below and, if you would like to use it, I can send you my R scripts.

(i) Small lakes in big landscape: Multi-scale drivers of littoral ecosystem in alpine lakes: ttps://www.researchgate.net/publication/281288501_Small_lakes_in_big_landscape_Multi-scale_drivers_of_littoral_ecosystem_in_alpine_lakes

(ii) Nodes in the alpine connectome. Exploring the linkages between riparian ecosystem and geo-climatic elements across the mountain environment: https://www.researchgate.net/publication/291344276_Nodes_in_the_alpine_connectome_Exploring_the_linkages_between_riparian_ecosystem_and_geo-climatic_elements_across_the_mountain_environment

Article Small lakes in big landscape: Multi-scale drivers of littora...

Article Riparian vegetation in the alpine connectome: Terrestrial-aq...

Hi

Article An Update on the Global Stressors and Constraints Affecting ...

Hello, there are several techniques to perform the ordination in biological communities. However, personally I like them for the information provided by self-organization maps. The use of neural networks in ecological modeling is very used nowadays due to the fascilities provided by the means of computation. The use of data mining is more recommended because of the complexity of ecological phenomena. This method is widely used because it is versatile in its use in non-linear and complex dats. There are several softwares to make them, however, in the R program you can find some packages that will help you to make them. I recommend you the following works:

Park, Y. S., Chon, T. S., Bae, M. J., Kim, D. H., & Lek, S. (2018). Multivariate Data Analysis by Means of Self-Organizing Maps. In Ecological Informatics (pp. 251-272). Springer, Cham.

Chon, T. S. (2011). Self-organizing maps applied to ecological sciences. Ecological Informatics, 6(1), 50-61.

Park, Y. S., Tison, J., Lek, S., Giraudel, J. L., Coste, M., & Delmas, F. (2006). Application of a self-organizing map to select representative species in multivariate analysis: a case study determining diatom distribution patterns across France. Ecological Informatics, 1(3), 247-257.

hello all, these are really good discussion for my study. If you have time, could you please mention how can I entry vegetation type(categorical variable) into discrete or continuous data types? because I don't know how to add this vegetation type data into my CCA data format requirement. Thanks.