Question on Ocean Vertical Velocity Data

[oyvindseland]

I am forwarding a question from Li Sirui Institute of Atmospheric Physics:
During my recent analysis, I observed that in some regions, the flow velocities reached extremely high values, such as 1e+8, which are evidently unrealistic for oceanic processes. This anomaly raises several questions about the data's accuracy and reliability.

Could you please advise on how to address this issue? Is there a known problem with the dataset, or could it be a result of some processing error on my part? Any guidance or recommendations you could provide would be greatly appreciated.

@matsbn

[LISIRUI1999]

My question mainly concerns whether the vertical velocity data, which can reach values as high as 1e+8, from the NorESM2-LM model in the CMIP6 piControl experiment on the 'gr' grid, is erroneous?

[TomasTorsvik]

@LISIRUI1999 , @oyvindseland , @matsbn , @JorgSchwinger
I believe this is a known problem with the dataset, rather than a processing error. The extreme vertical velocities coincide with seasonal breakdown of the stratification, especially for the Southern Ocean, causing large overturning cells in this region. This appears to be a by-product of the isopycnic vertical coordinate system applied by the NorESM ocean component, and is therefore present in both NorESM2-LM and NorESM2-MM datasets. We aim to replace the isopycnic vertical coordinate system for the next generation of NorESM.

[LISIRUI1999]

thank you for answering my question，I also notices that the model with gn grid does not have this question，It this mean the gn and the gr is different result rather than the different process， and also what is the seasonal breakdown of the stratification？Could you please tell me about it？I would be much appreciate.

[LISIRUI1999]

@TomasTorsvik

[TomasTorsvik]

[TomasTorsvik]

[TomasTorsvik]

@matsbn @oyvindseland @LISIRUI1999 @JorgSchwinger
Looks like I was mistaken regarding the "gr" version of the "wo" variable. It seems to correspond to the grid structure rather than the vertical velocity data, so it would seem the vertical re-gridding routine is doing something wrong here.