too cold to be true

[agustin@unex_es]

Hi all, I have ported CESM2.1.0 to a supercomputer in Extremadura (Spain). I have made a simulation of the 20th centure (1979-2015) but the results are completely discouraging with a very cold climate (I attach a plot of TREFHT given by the model an that of LEGATES, set 3 of the diagnostic package amwg_diag_20140804).

the compset used is:

2019-04-01 07:56:02: ./create_newcase --debug --case ../../../BHIST210_1x1 --compset BHIST --res f09_g17 --run-unsupported --mach login1 --compiler intel

the restart is

<entry id="RUN_REFCASE" value="b.e20.BHIST.f09_g17.20thC.297_01_v2">
<type>char</type>
<desc>
Reference case for hybrid or branch runs
</desc>
</entry>
<entry id="RUN_REFDATE" value="1979-01-01">

any help will be wellcome

cheers

Agustin

 

[erik]

A suggest you follow the guidelines in validating a port as described here:

7. Porting and validating CIME on a new platform — CIME master documentation

esmci.github.io

 

[agustin@unex_es]

Hi Erik, in first place I wish you a Merry Christmas (today is 24/12). After a big effort I could run the tests you suggest with a lot of 'FAILS' that I found them hard to understand. Find attached a file with the results. What I am most worried about is the difference between the results I get running CESM1 and CESM2. I attach a similar figure obtained with CESM1 and CESM2. If It was a machine problem why the results are so good with CESM1 and so bad with CESM2.

Thanks again,

 

[agustin@unex_es]

After my last post of 24 decembre, I has been working out the results of the test. A lot of FAILS come from the fact that the application cprnc was not compiled. However, after its compilation, there remains numerous FAILS that I have writen in the attached file. Any suggestions to solve them are wellcome.


Cheers,

Agustin

 

[agustin@unex_es]

I have made a new try using CESM2_1_3. I have change the netcdf version from 4.1.3 to 4.6.3 (C) and 4.4 (fortran). The problems follow the same. Attached is a figure of the daily mean temperature TREFH (2 m) at the point 12 N, 0W for the first year of simulation. As it may be seen the temperarure drops dramaticaly after the summer. I don't understand how a think like this may be happen. I will follow with the test as sugested by ERIK and change the post to the infrastructure section.

Thank very much

Agustin