weird methods: fit thousands of GAM, only look at those who converged, and then use Random Forest to see which variables are responsible for the fitted VPD-growth relation (like site conditions, forest structure)
VPD decreases growth, but even more the following year
don't observe a clear pattern on the iso/anisohydric continuum (only 7 species)
explain some positive VPD response by: "likely limited to sites where growth is enhanced by a temperature increase", "warmer temperatures associated with higher VPD could alleviate cold linitations of growth" (but maybe you can model it directly right? In our results can we make sense of the positive VPD response?)