In this study we are specifically interested in the role of convective momentum transport in a commercial fine-scale LES model that is used for wind predictions in the wind-energy sector. With this model, forced with the ECMWF IFS, a year-long of daily forecasts are run over Cabauw, the Netherlands, at a resolution of 40m in the horizontal and a resolution in the vertical of 8m that decreases with height to 80m. At this location atmospheric conditions range from stable to convective boundary layers, from clear sky to overcast days. The number of days with cumulus convection are surprisingly large: 144 out of 365 days.

Focusing on daytime hours, we separate days with and without cumulus convection, and with small and large cloudiness. For these different categories we show how the normalised momentum flux, and what this implies for drag in the lowest kilometre of the atmosphere. Here we are mindful of differences in background winds between the categories. A similar exercise is performed by separating cumulus days on surface buoyancy flux and cloud depth; four groups of increasing convection have been obtained.

It is found that in the presence of cumulus convection, the momentum transport is less linear than on overcast days. Clear-sky days show more drag in the lower half of the mixed layer compared to the cumulus case. Near cloud base, the drag is similar to that of cumulus days.

Separating cumulus days on convection we found that with decreasing buoyancy flux, the normalised momentum transport behaves less linear: in general cumulus days with the largest convection showed linear behaviour in the sub-cloud layer, cumulus days with the lowest convection showed a profile that tends to increase the wind speed near the surface and shows drag near cloud base.

Finally, using sensitivity experiments in which we remove the latent heating effect on buoyancy, we that the presence of moist convection specifically, change momentum mixing near and above cloud base, but not in the sub-cloud layer underneath.