You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
What default options do we want for the initial NWP configuration? #456
@barlage@dlawrenncar@martynpclark - can you please use this issue to list the options we want for the initial NWP configuration? My impression is that we're starting with clm5_0 but then changing some options. What I know so far is that we want SP with a few things changed:
I think it may be 2 PFTs, bare soil and the dominant PFT. We will need to
make a decision about the river model. This requires updated mapping
files, which we can have out-of-box for a CONUS configuration, but which
would need to be pulled into surface dataset generation workflow for the
more general runtime configurable grids that we will aim for eventually.
@dlawrenncar right now we don't have a tool chain where we can create the needed mapping files for river flow for regional grids. So I think it should be turned off, with the exception of those regional grids that we do the effort to create the mapping files for. I also think this will need to be implemented in small steps, first using an existing grid, then using the CONUS grid as a regional grid, and later fully adding the CONUS grid into cime. Once we get to that last step, we could probably enable the river model for the specific CONUS grid. You couldn't really do it easily out of the box, until that last step is reached. I got the impression from yesterday's meeting that river flow for NWP wasn't a critical thing especially to start with. I assume river flow WILL be critical for the "hyd" mode though. I think it would be better to start with a smaller set of requirements. Perhaps we should create another issue about the "hyd" mode?
From this mornings discussion two other settings are: oldfflag=1 and subgridflag=0. Right now those are only allowed to be set for Clm45, so that will need to change as well.
oldflag is essentially what's being used in NOAH-MP right now. So you want it as the out-of-the-box NWP option right now. And of course you want it to be an option that you can change. I also think that setting these means you no longer need STD_ELEV and SLOPE on the surface datasets, which is also good for NWP (since creating those maps takes so much CPU and memory).
Anyway, I think the bottom line is you want to be able to do both. To turn them on, and get them out of the box for NWP.
@ekluzek there are a lot of options in Noah-MP that we aren't going to support or use by default in our initial NWP configuration, which is why it isn't clear to me if we necessarily want these options initially. @barlage@dlawrenncar and/or @swensosc can you confirm what you want in terms of oldfflag and subgridflag for the initial NWP configuration?
I don't think we want the older version to be in the out-of-the-box NWP
configuration. We believe that the newer parameterization is an
improvement so it should be out-of-the-box. Users can revert to the older
version if they need to or want to for their research.
Activity
dlawrenncar commentedon Jul 27, 2018
ekluzek commentedon Jul 27, 2018
@dlawrenncar right now we don't have a tool chain where we can create the needed mapping files for river flow for regional grids. So I think it should be turned off, with the exception of those regional grids that we do the effort to create the mapping files for. I also think this will need to be implemented in small steps, first using an existing grid, then using the CONUS grid as a regional grid, and later fully adding the CONUS grid into cime. Once we get to that last step, we could probably enable the river model for the specific CONUS grid. You couldn't really do it easily out of the box, until that last step is reached. I got the impression from yesterday's meeting that river flow for NWP wasn't a critical thing especially to start with. I assume river flow WILL be critical for the "hyd" mode though. I think it would be better to start with a smaller set of requirements. Perhaps we should create another issue about the "hyd" mode?
billsacks commentedon Aug 9, 2018
Other things:
Note that Mike's dataset has no urban and lake
Turn off urban heat stress indices (via calc_human_stress_indices)
ekluzek commentedon Sep 6, 2018
From this mornings discussion two other settings are: oldfflag=1 and subgridflag=0. Right now those are only allowed to be set for Clm45, so that will need to change as well.
billsacks commentedon Sep 6, 2018
It wasn't clear to me: Do we actually want those for the initial, out-of-the-box NWP configuration, or just as options that could be turned on?
ekluzek commentedon Sep 6, 2018
oldflag is essentially what's being used in NOAH-MP right now. So you want it as the out-of-the-box NWP option right now. And of course you want it to be an option that you can change. I also think that setting these means you no longer need STD_ELEV and SLOPE on the surface datasets, which is also good for NWP (since creating those maps takes so much CPU and memory).
Anyway, I think the bottom line is you want to be able to do both. To turn them on, and get them out of the box for NWP.
billsacks commentedon Sep 6, 2018
@ekluzek there are a lot of options in Noah-MP that we aren't going to support or use by default in our initial NWP configuration, which is why it isn't clear to me if we necessarily want these options initially. @barlage @dlawrenncar and/or @swensosc can you confirm what you want in terms of oldfflag and subgridflag for the initial NWP configuration?
dlawrenncar commentedon Sep 6, 2018
billsacks commentedon Jan 11, 2019
From discussion with @barlage
We'll also reduce the number of canopy iterations (to 3)
No need to have a river model for now
We will support the NLDAS grid with the standard CLM5 configuration (so will have a standard surface dataset at that resolution).
We will not (for now) support the NWP option with any other grids (because we won't have the appropriate surface dataset)