I am looking at the influence of species proportions on nitrous oxide emissions using the diversity interaction model (Kirwan et al., 2009). I have six species - G1, G2, L1, L2, H1 and H2. However, H1 and H2 always have the same proportions which is introducing the collinearity. But I need to include H2 as biologically it influences nitrous oxide and I also can't transform it because then the species proportions don't add up to 1.

autoDI1