[CF-metadata] Question about cell_methods

From: Jonathan Gregory <j.m.gregory>
Date: Fri, 27 Jun 2014 16:03:15 +0100

Dear Jim

> I have a situation where there are multiple measurements within a 4D
> lon/lat/z/t volume, for which a single value is produced by
> calculating the weighted mean of all the measurements. This mean is
> of the form
>
> (sum over i of wi * vi) / (sum over i of wi)
>
> where wi is the inverse of the uncertainty in the measurement vi.
>
> So how do I represent that within cell_methods? Would it be
>
> lon lat z t: mean (uncertainty weighted)

That's the idea, assuming lon lat z t are the names of the four dimensions.
The format is actually "lon: lat: z: t: mean" (see 7.3.1). There is no
standardised way to record weighting, but it's fine to use a commment for that,
as you show.

Cheers

Jonathan
Received on Fri Jun 27 2014 - 09:03:15 BST