Hi Jon,
I suggest you look In David Warton?s new book, Eco-Stats: Data Analysis in Ecology (2023). He is one of the coauthors of ?mvabund.? Part III covers Multivariate Abundances, starting with Chapter 14. Exercise 14.1 is an example examining whether there are difference in the invertebrate communities in a revegetation study. Section 14.2 applies GEE (General Estimating Equations) for the multivariate invertebrate community data using the mvabund library (Code Box 14.1). In section 14.4, they do post-hoc univariate testing of the invertebrate community data, with adjusting P-values for multiple testing to control family-wise Type I error. So, perhaps this approach also will work for your data, also.
Good luck!
Best Regards,
Carl von Ende
cvonende at niu.edu<mailto:cvonende at niu.edu>