[CF-metadata] vertical coordinates and positive attributes

From: Jonathan Gregory <j.m.gregory>
Date: Wed, 29 Jan 2014 17:26:44 +0000

Dear Steve

> In practice I do not think that
> standard_name=depth and positive=up are necessarily in conflict (see
> bold text below). We fairly commonly encounter ocean model outputs
> in which the depths are encoded as negatives:
>
> 0
> -10
> -20
> ...
>
> There are merits to this encoding -- especially apparent in coupled
> ocean/atmosphere situations. (Also it preserves the right-handed
> coordinate system.) It would be quite peculiar (and a frequent
> source of error) to insist that an ocean model output file name its
> Z axis as "altitude", simply because it has encoded the depth values
> as negatives. The purpose of the 'positive' attribute is to inform
> the application when the encoding of values and the physical
> interpretation of the Z axis are reversed from one another.

Maybe it's annoying, but I can't agree with that. I would say that such depth
values are wrong. The standard_name depth is defined as "the vertical distance
below the surface". -10 m means 10 m above the surface. There's no point in
having standard names unless we use them correctly, and this is one of the
kinds of imprecision standard names can help eliminate. It's fine for the ocean
model to encode -10 m for 10 m below the surface if it wants to, but then it
should use the standard_name of height instead (or altitude, since distance
above the surface and above the geoid is approx the same thing in ocean areas).
In all the ocean model data I work with (mostly from CMIP), depth is a positive
number, as it should be.

Best wishes

Jonathan
Received on Wed Jan 29 2014 - 10:26:44 GMT