dtau directly at scenario's IR wavelength?

[abierjon]

Revision r2643 implements a dust particle size-dependent coefficient to convert the scenario absorption optical depth at 9.3µm to a visible extinction optical depth, before comparing it to the GCM visible opacity and inferring a dust injection rate.

It may be even more relevant to explore the possibility of a comparison done directly between 9.3µm absorptions, which would bypass the conversion (source of increased sensitivity and possible errors).

If so, one has to look in particular at the expression of the dust injection rate dustliftday in function of the opacity difference dtau, which will not be in visible anymore. Especially the coefficient alpha_lift, set up in initracer, must be updated by changing the Qext,VIS(reff_lift=3µm)=2.4 value in its expression by a function of Qext,IR(reff_lift) and omegaIR(reff_lift)
(with also some sensitivity studies on reff_lift)

[abierjon]

Actually, it would result in the same thing if the comparison was done directly in IRabs instead of VISext, because the ratio between tau_GCM_VISext and tau_GCM_IRabs would exactly be the conversion coefficient we use since r2643 to change the scenario.
Since people are more used to think tau in VISext, we will keep the comparison in the visible.

However, sensitivity studies on reff_lift can still be done and will require changes of Qext in initracer, which will be more detailed in a new proper ticket (#113)