F.A., M.P., G.L.S., M.R.B. and P.D.L. conceived the research project. N.L.G., P.D.L., R.T.O. and F.A. developed and tested the customised crucible-piston system to perform in situ moderate pressure, high-temperature X-ray microtomography experiments. F.A., M.P., G.L.S., N.L.G., E.W.L., R.A.B., R.T.O., D.D.G., M.E.H., H.M.M., R.A., P.D.L. and M.R.B. contributed to the beamline experiments. F.A. collected the volcanic rocks for the starting material. R.A.B. and H.M.M. prepared the starting material. F.A., M.P. and G.L.S. performed tomographic image reconstruction. F.A. and M.P. performed image processing and segmentation. F.A. performed image analysis. F.A. and E.W.L. performed rheological calculations. F.A. and G.L.S. performed numerical simulations using the dike model. F.A. performed chemical analysis. F.A. collected samples of the Etna 2001 eruption. F.A. acquired back-scattered electron images of Etna 2001 eruption samples. F.H. performed EBSD measurements. F.A., M.P. G.L.S., D.A.N., R.A.B., D.G. and M.R.B. wrote the manuscript, with contributions from all other authors.

Effect of defoliation on diversity and relative abundance of trophic groups. Effect size of defoliation on MOTU diversity (a) and relative abundance of reads (b) for each trophic group with 90% credible intervals. The barplot (c) shows the total number of reads (logarithmic scale) of each trophic group in the overall dataset. A multilevel linear model was fitted individually for each trophic group with a dummy variable for defoliation as predictor and a random factor accounting for the nested sampling design. MOTU diversity was standardized by the maximum value observed within each trophic group to obtain comparable effect sizes between groups. The colours correspond to the trophic class definitions (see Fig. 2).




Heidy Model Set 021