[CF-metadata] Two [simple] questions

From: Jonathan Gregory <j.m.gregory>
Date: Thu, 24 Jan 2019 14:34:56 +0000

Dear Lars

> 1. If I have (observed) daily precipitation with unit mm/day and want to store this in a CF compliant file, should I use standard name
> --- "lwe_thickness_of_precipitation_amount" with canonical unit metre and the "per day" part is inferred from the time bounds, or
> --- "lwe_precipitation_rate" having canonical unit "m s-1",
> --- if both are acceptable, is one to be preferred?

You can use whichever you prefer. They would also be distinguished by
cell_methods. The first is extensive in time, so it has cell_methods of sum,
the second is intensive and has cell_methods of mean.

> 2. The long_name is free, but should be descriptive. I am thinking of this as a succinct plot title, so the question is what practical limitation there might be in relation to different software. Is it advisable to use ASCII spaces (from
> Github issue #141 indicates that only standard ASCII character set should be used) ?

I expect others have something useful to say about that.

Best wishes

Jonathan
Received on Thu Jan 24 2019 - 07:34:56 GMT