[CF-metadata] x- and y-components

From: Jonathan Gregory <j.m.gregory>
Date: Mon, 17 Nov 2003 08:57:39 +0000

Dear Pieter and Brian

I agree with you too. This issue will affect all vector and tensor quantities.
It is inevitable that there will be two versions of many of them, because a
model might need to output such quantities (velocities etc.) both with respect
to its grid and with respect to latitude and longitude. Clearly they have to be
distinguished.

Therefore I retract my proposal to add eastward and northward sea water
velocity. Instead I agree that we should have x_sea_water_velocity and
y_sea_water_velocity. Also we should replace grid_eastward/northward_wind
with x/y_wind i.e. make the grid_ names into aliases. x/y components should
be used whenever x and y are not longitude and latitude. In most cases x and
y will be projection_x/y_coordinate in units of distance, but in the case of
the rotated spherical grid they are grid_longitude and grid_latitude in units
of degrees.

Best wishes

Jonathan

> > > The velocity components relative to the local x- and
> > > y-directions on the grid
> > > should be called grid_eastward_sea_water_velocity and
> > > grid_northward_sea_water_velocity (as for winds). These
> > > standard names are not in the table but I propose we should add them.
> >
> > East and north even if the grid direction is somewhere in between and
> > location dependent? (picture in attachment)
>
> I agree that using east and north in contexts where they don't apply is
> confusing. We have, for example, names like projection_x_coordinate, and
> projection_y_coordinate to identify projection coordinates. Perhaps we
> could use names like x_sea_water_velocity and y_sea_water_velocity with
> descriptions to indicate that the x velocity component is positive in the
> direction of increasing x-coordinate.
Received on Mon Nov 17 2003 - 01:57:39 GMT