Peter, think the analysis you mentioned are the usually used in the case of their informative pictured results. Here some paper on different kind of multi scale analysis

http://www.sciencedirect.com/science/article/pii/S0006320702003233

http://www.entsoc.org/PDF/MUVE/6_NewMethod_MANOVA1_2.pdf

http://www.pearsonhighered.com/assets/hip/gb/uploads/Mayers_IntroStatsSPSS_Ch14.pdf

http://www.umass.edu/landeco/pubs/cushman.mcgarigal.2002.pdf

http://www.stat.ncsu.edu/information/library/mimeo.archive/ISMS_1969_615.pdf

http://www.google.ru/url?sa=t&rct=j&q=&esrc=s&source=web&cd=5&cad=rja&uact=8&ved=0CD8QFjAE&url=http%3A%2F%2Fwww.datatheory.nl%2Fpdfs%2F89%2F89_06.pdf&ei=8A7IVIi6JIW2ygODzYHIDA&usg=AFQjCNGMxunSBJX7qFUzSP0wqwo9v9MNJQ&sig2=Gjuu-ouewaGiQwW_dBIKkQ&bvm=bv.84349003,d.bGQ

If you are just trying to figure out which of those variables are related to a single metric (such as biodiversity/eveness/total abundance/presence of a single species) re: your community, then any form of multiple linear regression (GLMs, or generalized additive models - GAMs) could also be used, as well as geographically-weighted regression. Many have used these to not only create relationships, but then also predict their response variable of interest across a spatial grid (raster). If you are interested in this let me know and I can provide more detail.

If you are truly interested in the community, then yes PCA/clustering analyses are your best bet. You might want to check out the 'vegan' package in R - it has a number of related analyses included detrended correspondence analysis, constrained correspondence analysis, similarity/dissimilarity indices, etc. Hope this helps.

Dear Nathan,

I would be interested in GLMs, GAMs and geographically-weighted models. Perhaps, during the work I can also compare them in term which of these are the best fitting for this type of ecosystem.

Please, contact me via private message!

Thank you in advance!

Peter