[Liwg-core] co-chairs notes 1-17-17

Bill Sacks sacks at ucar.edu
Tue Jan 17 15:40:47 MST 2017

Relevant notes below.

One key thing: JF suggests that the component diagnostics be run on the 126_rampCglc run so others can look at the impacts. Marika suggests that we point others to this output.


Coupled simulations

125: "turn off liqsupersat" actually means allowing supersaturation

127: Did a short run with just RF + 1-hour coupling + chlorophyl fix, because neither land nor Caspian Sea were ready yet 
Jim E: Is performance okay? Cecile and Jim will look

128: A problem was discovered with the new topography generation that was used in 126 – probably fine for troposphere, but problems with WACCM. So 128 has the same reworked Greenland topography as 126, but with the old topography generation software.
Julio has shown that the improvement in Greenland precip in 126 persists in 128 as well

129: New land = final physics configuration (mainly one change), plus new tuning

The latest CAM tag will give you the configuration of 125


Greenland update (Bill Lipscomb, with slides from Jan Lenaerts)

Total precip now looks very good

However, the run is now considerably colder, so that we now have substantially too low surface melt and runoff.

LIWG is looking at two things:

(1) Rain-snow repartitioning

Previously, we were using:

< 0 C  => snow only

> 2 C   => rain only

New plan:

< -2  => snow only

> 0  => rain only

Just making this change over glacier landunits. There is some physical justification, in that the lapse rate can be different over glacier surfaces.

(2) Changing bare ice albedo

Current value is about 0.5 over bare ice. Observations suggest values more like 0.4. Run underway to look at this.

This will just affect glacier landunits (mountain glaciers as well as ice sheets) – but NOT seasonally snow-covered areas.

JF suggests that diagnostics be run on the 126_rampCglc run so others can look at the impacts. Marika suggests that LIWG point others to these.

Bill L: Does anyone know why climate has cooled around Greenland? Recollection is that this was related to the changes Julio made. Bill L says that Jan saw little change from 125-126 in terms of Greenland melt, and saw bigger changes from 119-125 in that respect. Others don't know off-hand: there were a lot of tuning changes, etc. (including relvar change, which Andrew said was a high-latitude effect).

Dave B: Is RACMO dataset for 1850 or present-day? Bill L isn't sure, but thinks it's climatological 1960-1990, which could be different from 1850. So if we were a little low, that could be justifiable, but we're much too low.

Latest run shows some warming over Greenland, but that could just be variability, because the run is only short.

Keith L: As Gokhan showed, the Lab Sea is a bit too salty now, so there's some room for a little more runoff... as long as there isn't too much....

Dave L also looked at the albedo in land-only simulations, and sees positive albedo biases at edge of ice sheet – arguing for changing the bare ice albedo.


Update on land tuning (run 129)

Dave L: On the grid cell mean, the new land tuning is looking very good. However, there are two PFTs that are totally dead. They have figured out the problem with one, and are working on the other.

These changes probably won't change high latitude much (so it's probably okay to use 129 for looking at Greenland SMB)

Dave L would like to continue 129 further in the meantime, to make sure things generally look okay. Should turn on pft-level output to see how individual pfts are doing


Continuing runs

We'll continue 128 & 129

Should we start a run with Caspian sea? 
Some feel that we should resolve the G case problem, in case this is a problem that could affect B cases as well.
Dave L wonders if it's worth doing anyway. We could difference the runoff to see how this differs.


Meeting Friday

JF suggests we have a meeting Friday, probably 10 am
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cgd.ucar.edu/pipermail/liwg-core/attachments/20170117/97986ee6/attachment.html>

More information about the Liwg-core mailing list