Hi!I did not do the CAM5 runs, but I think what happened was that the prognostic dust (as described in Mahowald et al., 2006, Liu et al., 2013) was turned on, and thus dust emission was calculated prognostically within the model, allowed to vary, but there was no attempt to FORCE the dust emissions to match any particular trend (e.g. the historical trends from Mahowald et al., 2010) or account for changes in land use (e.g. as done in Ward et al., 2014). Thus, the time trends were from soil moisture or vegetation trends, wind changes or chagnes in atmospheric transport, and which tend to be quite small compared to what we see in the obs (as you can see in those papers).It is possible that the emissions were prescribed from the CAM4 runs (Lamarque et al., 2010), but my memory was that they were prognostically calculated. I do not think at this time that the ability to have prescribed aerosols was in the CAM5. If you need this verified, you should contact Steve Ghan or Phil Rasch, although Rich Neale or Cecile Hannay might know also when the prognostic aerosols were enabled for CAM5.
I do not know why they put Ds and MD in the table: there is only one of these. Maybe because also deposition impacts the land snow albedo, and thus someone else entereed it there, on accident? CAM5 dust uses calculations for emissions in the CLM dust module, to some extent, although modified by a soil erodibility map and sizes when it comes into the CAM5. Please note that the optics and the soil erodibility map were later improved for cam5: the ones in the cmip5 archive aren't very good: Albani et al., 2014, available from my web page.I would not reocmmend using the CMIP5 dust output from any of either CAM4 or CAM5 for science purposes, except to interpret CMIP5 results, because of the problems in the dust.If you have more specific quesitons, feel free to email me directly. ThanksNatalie
