Dear all,
I have encountered an issue when working with the vector products provided by ORAS5.
According to the ORAS5 overview tab, three groups of vector fields are provided on the ORAS5 model grid: ocean current velocity, sea ice velocity, and surface wind stress. Among these, only ocean current velocity is additionally provided as an eastward–northward rotated product.
For my research, I need the eastward–northward components of sea ice velocity and wind stress as well. Therefore, I derived a rotation matrix using the native and rotated grids of the ocean current velocity. When I apply this rotation matrix to the native ocean current velocity, the resulting eastward–northward components agree very well with the officially provided rotated ocean current product. The rotated wind stress also appears physically reasonable.
However, I obtain unexpected patterns when applying the same rotation to the sea ice velocity, particularly near the northern linking line connecting the two poles of the tripolar grid. For comparison, I examined the sea ice motion product from NSIDC-0116. Interestingly, the originally downloaded ORAS5 sea ice velocity already resembles the correctly rotated NSIDC product quite well, whereas the sea ice velocity after applying my rotation does not.
I further examined the grid definitions of the different vector variables and found that the native ocean current velocity and wind stress are defined on the staggered U/V grids, whereas the native sea ice velocity appears to be defined directly on the T grid.
This leads to my main question: Is the sea ice velocity provided by CDS already expressed in eastward–northward components, despite being described as a vector field on the ORAS5 model grid? If so, this could explain why applying the rotation matrix derived from the ocean current grid produces incorrect results for sea ice velocity.
If this is indeed the case, it may also be helpful to clarify this point in the ORAS5 documentation, particularly regarding the grid and vector orientation of the sea ice velocity product.
For reference, the figure below shows the monthly mean sea ice velocity for December 2016, comparing the different products/rotation approaches.
Could anyone familiar with the ORAS5 data processing or grid definitions please confirm whether my understanding is correct, or point out if I have overlooked something in the rotation procedure?
Thank you very much for your help!
Best regards,
Yupei