[RegCNET] Running RCPs using CMIP5 data
Graziano Giuliani
ggiulian at ictp.it
Tue Aug 13 11:23:45 CEST 2013
On 08/12/2013 05:56 PM, Syed Ahsan Ali wrote:
> Dear Graziano and all.
>
> I want to set up RegCM4.3.5.6 for running RCPs with latest CMIP5 data sets. The
> directory I have found for downloading CMIP5 data for RegCM is
>
> *http://clima-dods.ictp.it/data/d8/cordex/CMIP5/daily/*
>
> Is it correct? Secondly where is the supporting data for these RCPs i.e SSTs,
> Topography, Landuse and aerosols?
[...]
The input data are in http://clima-dods.ictp.it/data/d8/cordex
You need 6 hourly dataset on vertical levels to have the ICBC for the RegCM, and
they are in different directories for the different models we have written ICBC
interfaces.
The directory structure under cordex reflects the one on disk we use to run (the
inpglob in the namelist), so to download data you must respect same structure.
The CMIP5 directory contains instead data used in verifications and evaluation
of model characteristics and performances.
Assume for example you want to use HadGEM2 which is UKMO model to drive a
regional simulation.
0) Topography and Landuse data for RegCM are unchanged.
http://clima-dods.ictp.it/data/d8/cordex/SURFACE
1) Create the directory SST and HadGEM2
2) In SST download the data from
http://clima-dods.ictp.it/data/d8/cordex/SST
with HadGEM2 in the name:
ts_Amon_HadGEM2-ES************************.nc
3) In HadGEM2, recreate the directory structure:
HadGEM2 _
|-> fixed
|-> RF_______
| |-> hus
| |-> psl
| |-> ta
| |-> ua
| |-> va
|
|-> RCP85____
| |-> hus
| |-> psl
| |-> ta
| |-> ua
| |-> va
|
|-> RCP45____
|-> hus
|-> psl
|-> ta
|-> ua
|-> va
And download the data in the {hus,psl,ta,ua,va} directories, in case leaving out
the unwanted scenario (i.e. you want to run just the RCP8.5 and not the RCP4.5
scenario) and the unwanted period (i.e. look at the dates in the filenames to
exclude data).
Then point inpglob to the directory where {HadGEM2,SST} directories are in.
Other option is to directly use the ESG federation servers to download six
hourly dataset: note that we have downloaded just the first release (r1i1p1),
while some models have already uploaded to the ESG servers new releases of the
datasets fixing bugs or updating the model code.
In this case, you may probably want to have a look at the ICBC code to modify
the file names built at run time by the sst/icbc programs.
Another note, we have interfaces ready for only some of the CMIP5 models (the
ones in the model code), and albeit we plan to develop in the future a more
general pre-processing to avoid re-coding on every GCM model name/content/quirk
change, as of now we have other work to do with higher priority, and you are
left with fortran code editing :( .
Third, the download and coding of different interfaces has been started by
different users here at ICTP, so each model may need data placed in different
directory structure. The directory structure on:
http://clima-dods.ictp.it/data/d8/cordex
is anyway the one we use on disk, so replicate it and you should have no problem.
For Aerosol datasets, you must download for the BC the data in the
http://clima-dods.ictp.it/data/d8/cordex/AERGLOB
directory. Again, you may want to download only the scenario you need.
Graziano.
P.S.: Some dataset are still downloading and may not be complete (CNRM RCP8.5 is
giving us headaches !!)
--
Graziano Giuliani - Earth System Physics Section
The Abdus Salam International Centre for Theoretical Physics
Strada Costiera, 11 - I - 34151 Trieste Italy
More information about the RegCNET
mailing list