overturning stream function standard names

[taylor13]

Question

For several "overturning stream function" variables, two similar options for standard_name are available. As an example, consider ocean_meridional_overturning_mass_streamfunction and ocean_y_overturning_mass_streamfunction. Both of these must be functions of some vertical coordinate, but when the horizontal coordinate is not latitude, the second apparently applies. There currently are no names involving overturning stream functions applying to the east-west direction.

I do not think standard names should distinguish between quantities that are functions of particular coordinates. The dimensions (and perhaps cell_methods?) can serve to define the dimensions that a quantity is a function of. So, my question is, should we have avoided defining the similar two standard names, and simply defined ocean_meridional_mass_streamfunction? The dimensions would be different in the two cases ("latitude, depth" vs. "grid_latitude, depth" for a rotated pole coordinate system), but they would share the same standard_name.

Curious what others think about the general principle that we avoid differentiating among coordinate systems in standard names.

[JonathanGregory]

Dear Karl

I think that meridional and y in these names is included to indicate the direction of a vector. Even though these are scalar fields, they represent velocities, in the latitude–depth and y–depth planes respectively. The word overturning signals the "depth" aspect. Only one dimension is freely chosen for a 2D streamfunction, so we could name them by the orthogonal direction, I suppose, i.e. "zonal streamfunction" and "x streamfunction", but that would be very obscure, not self-explanatory! I don't think the primary purpose of these words is to indicate the grid.

Best wishes

Jonathan

[taylor13]

Hi Jonathan,

Two of the three velocity components for ocean flow can derived from the 2D streamfunction, and your argument is that the direction of those velocities should be indicated somehow in the standard_name. Unlike the u and v components of horizontal velocity, however, the directions of the 2 velocity components derived from the streamfunction can be inferred directly by looking at the dimensions. If the dimensions include depth and latitude, then it is the vertical and north-south velocities that can be calculated. If the dimensions include depth and y, then it is the vertical velocity and the horizontal velocity in the y direction than can be calculated.

And the ocean barotropic streamfunction would be a function of longitude and latitude (or, more generally, x and y), but not depth, enabling one to calculate transport in the horizontal directions indicated by the dimensions.

So, there really is no need to define separate names for all these streamfunctions; they would all be included in ocean_mass_stream_function.

I do, however, understand that users might not want to look at the dimensions, so including an indication of the "plane" of the streamfunction in the standard name as we have done doesn't hurt. And I think we agree that grid information should normally not be included in the standard names, so I'll close this discussion.

thanks for your thoughts,
Karl