Closed penguian closed 2 years ago
@jxs599@nci.org.au edited the issue description
@jxs599@nci.org.au changed status from new
to closed
@jxs599@nci.org.au set resolution to fixed
@jxs599@nci.org.au changed milestone from 3. Implementation
to 1. Closed
@jxs599@nci.org.au commented
pushed to trunk
@ccc561@nci.org.au set keywords to nogit
keyword_nogit
resolution_fixed
type_code improvement
| by srb001@csiro.auAdd standard trunk@HEAD vns of "sumbal" values across 3 single-site forcing. Where sumbal is the combined cell average, cumulative across time steps, water and energy balance. i.e.
It has been found consistently that this quantity is very sensitive to any (inadvertent) change in a model "process". As such it is frequently used as a test of bit reproducability when making changes to the model that should not change the results.
Currently this is not set up for the MPI version and available only for the serial version. As this is somewhat limited by several factors:
we broadly address this by testing at 3 different sites that hopefully at least capture a good range of PFTs and climate. It will be necessary to manually test across a range of configs. Testing globally using GSWP2 forcing is also set up and manual checking of values in the output (log_cable(_1986).txt is required. Configs are limited (out of the box) by those set up in:
https://svnserv.csiro.au/svn/CABLE/dev/users/jxs599/GenericCABLETicketDev/
These will not necessarily be updated with every trunk revision. However svn should be able to clarify at what revision a baseline sumbal file was updated.
The 3 flux tower sites are at Tumbarumba, Ampero and Hytiallia. The corresponding baseline sumbal files are respectively:
CABLE runs in the testing config at least write out a file for comparison,
new_sumbal
For MPI, global runs using GSWP2 forcing. The corresponding baseline file is:
CABLE runs by default write out a file for comparison,
log_cable_1986_1.txt
Issue migrated from trac:318 at 2023-11-27 11:40:03 +1100