0

I have previously had success rotating my NMDS plot with reference to an external variable via this package (https://rdrr.io/rforge/vegan/man/MDSrotate.html). However, with my current data, there is no external variable that I am able to rotate the ordination in reference to in order to have most variation occur along the x-axis NMDS1. Ideally, the ordination lines (blue, red, orange, cyan) would be nearly horizontal. Is this possible to achieve mathematically? How would I do this?

NMDS plot with environmental variables as vector overlay

`

## My code
D_glt_CHEMspe   <- vegdist(glt_CHEMspe, 'bray', na.rm = TRUE) # use a dissimilarity measure of your choice
(ord_glt_CHEMspe <- vegan::metaMDS(D_glt_CHEMspe, k=2, maxit=500, try=500, weakties=F))
plot(ord_glt_CHEMspe, type='t')
plot(ef_glt_CHEMspe <- envfit(ord_glt_CHEMspe, env[,'Year'], na.rm = TRUE))
ef_glt_CHEMspe
ord_glt_CHEMspe <- vegan::MDSrotate(ord_glt_CHEMspe, env[,'Year'], na.rm = TRUE)

plot(ord_glt_CHEMspe, type='t')

`

It rotates along a predefined variable as a transposed linear gradient, but I would like to rotate along potentially a numerically defined degree (ex: 85 degrees clockwise).

dijon_djinn
  • 11
  • 1
  • Check rotation matrix https://en.wikipedia.org/wiki/Rotation_matrix. You need to fill the matrix with cos and sin of the desired rotation angle. – Jari Oksanen May 28 '23 at 12:15

0 Answers0