Problems with optical depth in the 2-layer aerosol scheme

[jvatant]

In the 2-layer aerosol scheme, there was a problem with the renormalization of stratospheric and tropospheric layers giving almost all the opacity to the tropospheric layer.

This normalization was corrected in r2254 but it then leads to too much extra opacity in case of an intermediate layer.

To fix this and also add more flexibility (eg having different optical properties for the different layers), we will move to a n-layer scheme with aerosol layers distinct the ones from the others.

[jvatant]

Fixed by r2297