The ms is of very interest in terms of examining new approaches for the improvement of the oil spill predictions due to surface forcing uncertainties .

The use of the ECMWF ensemble wind forecast can provide valuable information to the response agencies regarding the  impacted area, both at sea and at shoreline, as well as of the variation of the weathering processes.

Minor comments:

Clarify the temporal resolution of the used ECMWF deterministic and ensemble wind  forecasting data,  as from the Figure 4 it seems that are 3 hourly? Similarly, provide the temporal resolution of the used CMEMS Med MFC sea currents.

Clarify if the DA is applied also to the southern part of the convex hull area, as from Figure 5 the DA it seems that it was applied only in the northern part.

The figure 8 to be renamed as figure 8a and add an additional map as figure 8b showing the extend of the deterministic and ensemble spill extend on the sea surface and on shoreline only. The transport of the dispersed oil, i.e. the transport of the subsurface oil in oil spill modeling is calculated using sea currents, not  winds.

It will be an added value of the ms the provision of the plots of the main  weathering parameters derived from the  deterministic oil spill simulation and those derived from the ensemble oil spill simulations (mean averaged).

For operational oil spill predictions is of interest to the response agencies to provide the  run time required for the deterministic oil spill prediction  and of the run time required for the ensemble oil spill predictions for 48 hours, 72hours, 120hours and 240hours.

In the present manuscript the Impact of using an ensamble atmospheric forcing on a oil trajectory and wheathering model is studied.

I Found the paper well written and focused on a relevant argument on which it shed some light.

The paper does not clarify how much the approach can improve the solution, while it observes an increase of possible oil beaching (20 to 100 percent more than the deterministic solution). Anyway, i found that the paper deserves to be published.

I just suggest an inmprovement of section 2 with a more detailed description of the differences in the implementation of the ensamble vs deterministic simulation. In particular it is not clear to me the approach used in simulating with the ensamble solution. Is it used the "ensamble" averaged solution or the members of the ensamble are treated as single runs? In other words does the oil spill model is ran 50 times and then actually an ensamble oil spill trajectory and evolution of oil is considered? By reading "ensamble oil spill model" I would be induced to figure out an actual ensamble of trajectory, but it is unclear to me if Authors actually performed an ensamble of trajectory. In negative case, i.e. if authors just ran a "deterministic" oil spill by using the averaged solution of an ensamble atmospheric forcing, I would suggest to revise the text rewording sentences relative to "oil spill ensamble".