"Wrong" or not the ideal map projection

[Alex_Cantrel]

Dear forum.

I accidentally ran a long-period simulation using the 'Mercator' projection for a 3-nested domain (9km, 3km, 1km) case over the North Sea.
I know that the suggested projection for this area is 'Lambert'.
So I conducted a weekly validation run and there are discrepancies between the two different projection runs (i.e. 'Mercator' vs 'Lambert'), as expected.
I am not sure if this would make sense, however, I was wondering if there is any possibility to post-process the Mercator-executed WRF runs and project them to a Lambert map to avoid proceeding with the long-period run from scratch?
How could I confirm that the 'Mercator' runs are wrong?
Thank you very much in advance.

 

[Ming Chen]

Did you see large differences between results on Mercator and Mercator projections? I don't think map projection can affect the accuracy of model simulation. It only affects the distortion degree of model grids ( represented by map factor in WRF) and thus the time step may need to be adjusted.

For your case, if the model runs to the end successfully, I wouldn't worry for the results even if the map projection is not appropriate.

Remapping the results to lambert projection won't improve the accuracy.

 

[Alex_Cantrel]

Thank you very much for your reply.

The model ran successfully. However, yes, I can detect discrepancies (e.g. in wind speed profiles at specific time steps), especially in the lower boundary layer (e.g. <1000m height).

I understand that remapping won't improve the accuracy.

However, I wonder, if I continue the year-long simulation switching to Lambert could the two different projection results be combined?
Furthermore, I don't know if this is helpful. However, the MAPFAC_M for the Mercator case has a min_value = 0.9048, and a max_value = 0.9633.
While, for the Lambert case I get: min_value = 0.9990, and a max_value = 0.9995, which makes more sense.


Thank you again for your help.