geoschem / geos-chem

GEOS-Chem "Science Codebase" repository. Contains GEOS-Chem science routines, run directory generation scripts, and interface code. This repository is used as a submodule within the GCClassic and GCHP wrappers, as well as in other modeling contexts (external ESMs).
http://geos-chem.org
Other
170 stars 165 forks source link

Generalize subsetting of HEMCO input fields w/ additional indices, e.g. 4D input data, xy+"ensemble=3:extradim=2" #2570

Open aschuh opened 6 days ago

aschuh commented 6 days ago

Your name

Andrew Schuh

Your affiliation

Colorado State University

Provide a clear and concise overview of the new feature requested.

How difficult would it be to allow the parsing of the additional constraint on the "spatial dim" field in the HEMCO config? I'm increasing running ensemble sims w/ GEOSChem and having the ability to more arbitrarily subset larger dimension files would be really really useful.

For example, I can run emissions from variables like the below by adding in "PFT=$PFT" constraint.

float RESP.BAL(PFT, time, lat, lon) ;
    RESP.BAL:units = "umol m-2 sec-1" ;
    RESP.BAL:_FillValue = -1.f ;
    RESP.BAL:long_name = "RESP balanced flux" ;
float ASSIM(PFT, time, lat, lon) ;
    ASSIM:units = "umol m-2 sec-1" ;
    ASSIM:_FillValue = -1.f ;
    ASSIM:long_name = "ASSIM Flux" ;

It would be nice to go one step further and be able to parse that constraint so we can run hyper sliced emissions from something like:

float ASSIM(ARBVAR,PFT, time, lat, lon) ;

with "ARBVAR=1:PFT=2" type strings.

Thoughts? The sooner I can know whether this is possible (and/or easy), the better. I'm constructing a large code base that will have to either split out a bunch of emission files or subset from them on the fly (which is much better).

thanks, andrew

Reference(s)

No response

Will you be implementing this feature yourself?

No

Additional information

Not sure if this goes to GEOSChem or HEMCO coders .... or how long it might take to implement.

yantosca commented 10 hours ago

Thanks for writing @aschuh and thanks for your patience. I've been traveling in these days and have just gotten back.

I believe @christophkeller was the one who wrote the ensemble parsing feature into HEMCO. I've never used HEMCO with ensemble data myself.