[MITgcm-devel] Re: NetCDF and interpolation of forcing data

[Dimitris Menemenlis]

Hi Ed,

> INPUT:  Do you have any need/use for an MITgcm interface for 
>    reading and interpolating NetCDF data?

None of the input files I use are in netcdf format.

>   If so, what do you want it to do and/or look like?

If I were to use netcdf input files I would want them to contain grid
information (latitude and longitude), time stamps, repeat period, name
of field, units, and sign information.  In other words all the
information that is now contained in data.exf and data.exf_clim files.

>   Are the routines in pkg/exf 
>    sufficient or would you like to see a different approach? 

pkg/exf is not yet complete or as clean as it could be, but it's
getting there and it has most of the functionality that is needed.
Eventually it would be nice to combine exf and exf_clim and to provide
separate repeat_period capability for each field, but it's not
high priority.

>  OUTPUT:  Do you have any needs in terms of MITgcm NetCDF 
>    output--either the actual model output or post-processing?
>    And if so, any hints on how you'd like it to be done?

The minimum needed would be grid information, time stamps, units and
sign convention, and output field name.  Also you should allow for
eventual inclusion of output filters, e.g., interpolation, vector
rotation, or bin-averages on arbitrary output fields.
... and it has to be fast!

D.

-- 
Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
Jet Propulsion Lab, California Insitute of Technology
MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
tel: 818-354-1656;  fax: 818-393-6720