-
Notifications
You must be signed in to change notification settings - Fork 23
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Change JRA55-do data directory #120
Comments
Should this be the JRA55-do path used for the default "vanilla" control directory repos? Also @nichannah - will YATM require modification to handle these filenames? Or should we just make a lot of symlinks with names in the format we want? |
Yes.
Not that I know of.
I was in the middle of testing this when other things cropped up. As a work-around for supporting the frankly annoying filenames I did as you suggested and made a bunch of links:
It ain't pretty but it works. The current status is that it ran ok but did not reproduce results using the current JRA55-do dataset. This is problematic. |
From Simon's email today it seems we should move to JRA55-do v1.4.0 but I can only find 1.3 in |
It will need to be downloaded, but there is a semi-automated process for that. I will consult with Paola. |
Thanks @aidanheerdegen |
I've lodged a help ticket at NCI to request download of 1.4.0. |
NCI has downloaded updates to JRA55-do here |
The naming convention has changed for 1.4.0:
|
Atmos data is more relevant:
|
the time component of the filename is not the same for all files
|
for comparison, here's
|
I think it will cover all cases to have a new key |
the month and time part of the filename can be specified in forcing.json which I expect (but haven't checked) will be the same for all files of a given field in a given version of JRA55-do, except perhaps for 2019 which is of course incomplete |
make_remap_weights.py and make_remap_weights.sh specify |
scratch that thought - they're RYF files |
I've tested PR COSIMA/libaccessom2#26 with these runs
using these scripts
Runs 1 and 2 are identical, as expected. @aidanheerdegen also noted above that This is because the near-surface temperature and humidity ( I've also confirmed that the month and time part of the filename is identical for all years (except the final incomplete year) in each variable we use in So in summary I'm confident that PR COSIMA/libaccessom2#26 is ready to use. |
Remaining things to do in all IAF JRA55-do 1.3 configs:
These have been done in COSIMA/1deg_jra55_iaf@283e0bc but not at the other resolutions. |
Closing. Both items above have now been done in the gadi-transition branch of all RYF and IAF configs (not including the minimal 0.1 deg configs). |
The JRA55-do atmospheric dataset is currently used to force the ACCESS-OM2 models.
At the moment NCI users access this data in the
ua8
directory where CMS (Paola) is downloading and updating the data.JRA55-do is now an official dataset of input4MIPS, which is part of CMIP6. This means it is available under the
qv56
project and centrally managed by NCI.This is a good thing. It will reduce Paola's workload, be downloaded automatically, and most importantly will be versioned, which was something that we discussed at TWG and definitely wanted.
Anyone who wishes to use this dataset will have to join
qv56
, which means anyone who wishes to run the ACCESS-OM2 models at NCI will also have to join.We will need to determine the file paths. The data is here:
/g/data/qv56/replicas/input4MIPs/CMIP6/OMIP/MRI/MRI-JRA55-do-1-3/
So for example:
This will need to be tested. I'll get around to it, but if someone else wants to put up their hands feel free and feedback to this issue so work is not duplicated.
The text was updated successfully, but these errors were encountered: