Some very valid qualms were raised about doing a straight Bray-Curtis dissimilarity timeseries analysis a la Bagchi et al 2017.
Idea from Morgan/Ethan: do a pseudo-wavelet thing, calculate dissimilarity between all pairs of points one year apart and plot. Same for all pairs of points 2 years apart, and so on. If a shift is really major on a short and long time scale, it will come out as a large dissimilarity on multiple measures.
Some very valid qualms were raised about doing a straight Bray-Curtis dissimilarity timeseries analysis a la Bagchi et al 2017. Idea from Morgan/Ethan: do a pseudo-wavelet thing, calculate dissimilarity between all pairs of points one year apart and plot. Same for all pairs of points 2 years apart, and so on. If a shift is really major on a short and long time scale, it will come out as a large dissimilarity on multiple measures.