Here are now the final results on the simulations regarding the influence of the recoil mass of the PR, SR, and BS mirrors on the light scattering and the strain noise of KAGRA.
Final means that I will use these data for the publication which I am now writing.

I distinguished two different cases.
First, I assume a recoil mass made of a perfect lambertian scatterer. And second is to run the simulations with the BRDF data of titanium (roughly polished) which will be closest to reality.

The two last pictures are for BS, backside.

It should be noted that the data in the figures are given in 1/(sqrt(Hz) sr) and have to be multiplied by the respective solid-angle toward the beams waist to be comparable with KAGRAs goal sensitivity!
This will further decrease the values (by a magnitude of 10, or so), so I decided to leave the data as they are to keep a good overview of their relation to KAGRAs goal sensitivity.
 

It can be seen that even for the BS, we only have a neglible effect on the strain noise due to the scattering on the recoil mass, which is of course a good sign.