Hello Jonathan, Karl many thanks for all the feedback; It looks like I will need to get to work with a package of standard_name definitions. I had hoped to circumvent this effort, but if you think that is not possible, then it will need to be done.
I will report back when we have collated a set which meet our immediate needs mark ________________________________ From: CF-metadata [[email protected]] on behalf of Karl Taylor [[email protected]] Sent: 30 July 2014 22:09 To: [email protected] Subject: Re: [CF-metadata] volume integral Dear Mark, I agree with Jonathan, that it is a different quantity and needs a different standard name. Since it is extensive, the default (and correct) cell_methods is "sum". Although no longer a function of lat, lon, and depth, you would probably want to define those as (scalar) coordinate variables (with the coordinates attribute) and include the "bounds" for each of the dimensions. Additional information/options can be found in the CF conventions sections 7.3 and 7.3.4. You probably already considered all of this, but thought I should make sure. cheers, Karl On 7/30/14, 7:04 AM, Jonathan Gregory wrote: Dear Mark We have a requirement to calculate volume integrals for a large range of model diagnostics. We would like a method to correctly identify these derived fields. I am interested in creating a new cell_method, an integral, which would allow us to use this approach with any standard name. Such a cell method will necessarily alter the units of the quantity by a factor dependent on the dimensions integrated over. For example, I would like to store a data variable: double deltaSeaIron(time, depth, lat, lon) deltaSeaIron:standard_name = "tendency_of_mole_concentration_of_iron_in_sea_water_due_to_biological_production" ; deltaSeaIron:units = "mol s-1" ; deltaSeaIron:cell_methods = "depth: lat: lon: integral" ; I can appreciate why you propose that but I would say that it is not consistent with the intention of cell_methods or the practice in the standard_name table. cell_methods is for describing variation within cells. Doing an integral is a change of variable. It converts an intensive quantity into an extensive quantity. There are very many other examples of this - hundreds, I expect. For example, rainfall_amount is a time-integral of rainfall_rate, and they are described by different standard_names, not distinguished by cell_methods. atmosphere_mole_content_of_ozone (mol m-2) is the vertical integral in the atmosphere of mole_concentration_of_ozone_in_air (mol m-3), again distinguished by standard names. In the case you mention, I think a new standard_name is needed, something like tendency_of_ocean_moles_of_iron_due_to_biological_production A construction like that would be analogous to many existing standard names (for many different species) of the form tendency_of_atmosphere_moles_of_X. Best wishes Jonathan _______________________________________________ CF-metadata mailing list [email protected]<mailto:[email protected]> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
_______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
