We used purely radial LDI simulations, so the “clumps” observed in the simulations actually represent thin shells of material flowing out from the star. There is reason to suspect that adding more dimensions would affect clumping, since Rayleigh-Taylor or thin-shell instabilities would break up these spherical shells of material. However, preliminary 2D models produce very similar radial variations in clumping factor, so we stayed with the 1D models.

It is interesting to note that despite the variation in metal content across our models, there is some convergence of clumping factor at large radii. Previous simulation work (Dessart and Owocki 2004) suggests that the outer wind represents a purely gas-dynamical regime, with the LDI substantially weakened by the lower densities. The metal content should have little effect in this region of the wind. Dessart and Owocki propose a mean clumping factor of roughly 4 in the outer wind (going to thousands of stellar radii). Any power law exponent derived from radio diagnostics would then be unaffected by clumping.