Hi!
I have been working with cave invertebrates from semi-arid regions, and have observed some interesting responses from the community to variables at the local and regional level. I know which variables explain their patterns of similarity and richness at different spatial scales. However, I would like some way to be able to illustrate the occupations of the different groups according to the environmental gradients, especially to verify that the same groups maintain the same occupations. Is there a way to test and illustrate this for the whole community? I checked some works that do this according to the overlapping of the functional traits of the groups. However, I would like to do this according to the occurrence of the groups according to the environmental gradients. I appreciate any possible help.
You might find ordination methods helpful for this. I used different methods for my studies on birds and spiders during the years. Check figure 3 in our latest work (Domer et al. 2019) on bird communities in riparian corridors. In this case we used nMDS, although Primer 6 can offer many other ordination types (you can decide whether you want to ordinate plots by species or species by plots). For species composition, similarities are compared among samples (all pairwise combinations) using either Jaccard or Bray-curtis index. It will then plot all your samples on 2 (or 3) axes state-space. If you had 3-4 different "treatments" you can then mark them differently and see how the species cluster. Any environmental gradient that you have measured can also be entered to the model, and show as an arrow (vector) that starts at the axes origin. Its direction and length will be associated with the species cluster in the same direction, etc.
You might find ordination methods helpful for this. I used different methods for my studies on birds and spiders during the years. Check figure 3 in our latest work (Domer et al. 2019) on bird communities in riparian corridors. In this case we used nMDS, although Primer 6 can offer many other ordination types (you can decide whether you want to ordinate plots by species or species by plots). For species composition, similarities are compared among samples (all pairwise combinations) using either Jaccard or Bray-curtis index. It will then plot all your samples on 2 (or 3) axes state-space. If you had 3-4 different "treatments" you can then mark them differently and see how the species cluster. Any environmental gradient that you have measured can also be entered to the model, and show as an arrow (vector) that starts at the axes origin. Its direction and length will be associated with the species cluster in the same direction, etc.
Canonical ordination
https://esajournals.onlinelibrary.wiley.com/doi/10.1890/0012-9658%282003%29084%5B0511%3ACAOPCA%5D2.0.CO%3B2
You may want to check out this book "Numerical ecology with R" from Borcard 2011. It has R scripts for multivariate analysis which will provide you numerous ways to illustrate your results.
I also use different forms of ordination to show responses to environmental gradients (PCA, CA and CCA and RDA). A power tool is to use variation partitioning (pRDA, pCCA) to determine unique and shared effects. A trial version of Canoco used to be available or a good alternative is to use R.
Hi ! I could add yo the previous responses the use of "envfit function" to show in a same graphic the ordinations plus their relation to the most explanatory environmental variables (I used in Pero et al. 2020, in STOTEN for benthic invertebrates). In addition, rank-abundance curves could be used to show each taxa distribution within the different comminities or assemblages (I used in Pero et al. 2019 Freshwater Science). I hope that this could be usefull for you. Good luck!
- Badges
- Science topic