Hi @chenhui.jin, just to chime in (mostly because this is a very interesting discussion and I want to follow it), it’s highly unlikely that the model will be happy to ingest ancillaries with NaNs in them.
1 Like
Watch out for boundary condition issues if your regional domain land masses don’t match the driving data.
2 Likes
Hey Paul,
Just reporting back, I talked to @sonyafiddes after I chatted with you about running rAM3 for a high latitude case, but it seems like rAM3 (at least currently) won’t work for my case. The main issue (aside from the high latitude problem) is that for the simulations I plan to run I need coupled aerosol and cloud microphysics (GLOMAP-mode and CASIM) which I believe isn’t yet coupled in rAM3…the other issue is that driving the model with ERA5 won’t have the aerosol fields. Jason and Sonya have a UM configuration setup with coupled aerosol and cloud microphysics, driven by a global UM run which has the aerosol fields so I will most likely be using that.
Is there any plans to couple CASIM and GLOMAP-mode in rAM3 that you know of?
Hi @alanah.chapman,
I haven’t heard of any plans to couple CASIM and GLOMAP-mode in rAM3.
Do you know if any other group is doing it? i.e. another group in the Momentum Partnership??
It might be worth talking to the scientists first, then raising it with ACCESS-NRI for discussion in workplans (if it is possible).
@sonyafiddes have you heard anything from other groups?
Hi Team,
We are running a similar version to the suite as ram3 (RAL3.2), but like Alanah said - we have coupled this to GLOMAP. Other groups in the partnership have done this (Hamish Gordon helped us, aka Matt, set it up). We have had this working for sometime now 
The biggest differences to ram3 will be we are driving with the UM because we need the aerosol emissions (like Alanah said), and we have to turn off a few of the land-surface things (like leaf area index) to run for Antarctica - as you probbaly know - most of those ancillary files only go to 60S, we run with a different orography (see Jason’s pre-print here - note this is for a version much more similar to ram3 - without coupled arosols). And I cant remember if ram3 has CABLE? if it does, our version doenst (but again - Antarctica so it doesnt really matter).
Hope this helps
https://essopenarchive.org/doi/pdf/10.22541/essoar.175339123.30294645
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2024JD042311
Thanks @sonyafiddes. I guess the question is whether the community needs an ACCESS-NRI supported example of the RNS running with AM3 initial conditions and lateral boundary conditions (and CASIM and GLOMAP). Perhaps it is worth you and @alanah.chapman talking with the Atmosphere Working Group or suggesting the requirement for the ACCESS-NRI Workplan?
1 Like
By the way, ACCESS-rAM3 can run from UM start dumps but I need to commit fixes to logic in the rose/cylc introduced when I added in the local changes to get the BARRA replacement etc to work. Please reach out if you need access to it and I will get back to you.
2 Likes
<Region>_<resolution>_ancil_lct tasks are completed. I proceed to modify the land-sea masks.
I was just wondering if I should only modify qrparm.mask_cci, which is linked to qrparm.mask, or I should modify all the qrparm. files
Here is my modification:
mask_era5_dir = '/scratch/if69/cj0591/cylc-run/u-bu503/share/data/ancils/IndianOcean_Plus/era5/qrparm.mask_cci'
mask_era5 = iris.load_cube(mask_era5_dir, 'land_binary_mask')
lat = mask_era5.coord('latitude').points
mask_era5.data[lat < -60, :] = 0
Should I output the modified data into netcdf format? I had some issues when I try to save the data using iris.save
iris.save(mask_era5,'/scratch/if69/cj0591/cylc-run/u-bu503/share/data/ancils/IndianOcean_Plus/era5/qrparm.mask_cci_modified')
ValueError: Cannot save; no saver
1 Like
Hi Chermelle,
That’s a good idea - there are quite a few users of this configuration (5 at my count) so having it centrally support would be amazing, and running with ACCESS-AM3 not the UM. I can suggest it.
1 Like
Hi @chenhui.jin,
To modify the qrparm.mask you need to use mule.
Mule can be a bit challenging.
A fairly complicated example of using mule can be found on our GitHub. I can DM you if you want.
The stash code of the land mask is m01s00i030 so you would be searching for f.lbuser4 == 30 to replace the land sea mask which you could write the code to edit or read in a replacement. Please reach out if you need help.
Hi @cbengel,
Thanks for the heads up
I am currently trying to use ants to modify the qrparm.mask following this Ancillary tools — 21st Century Weather Software Wiki
If I have no luck on this, I would reach out to you regarding mule
1 Like