Problem with the leap calendar in ORCHIDEE
The problem of incorrect interpretation of the forcing with the Gregorian calendar by ORCHIDEE has been discovered.
It is likely connected to the recent corrections done in the ticket:
https://forge.ipsl.jussieu.fr/orchidee/ticket/433
The CRUHARv3.2 forcing (produced within the VERIFY project over European region at 0.125°)
with the time axis defined as shown below and containing the day of February 29 is not read correctly
in ORC22 rev7330 and ORC3 rev7267 using the old driver. The graph below for one pixel close to Paris location
shows that it was working fine in the previous model revisions.
float time(tstep) ;
time:units = "seconds since 1901-01-01 00:00:00" ;
time:long_name = "Time axis" ;
time:calendar = "gregorian" ;
|
The change of the calendar to the noleap-format and excluding the data for February 29 allows avoiding the problem:
float time(tstep) ;
time:units = "seconds since 1901-01-01 00:00:00" ;
time:long_name = "Time axis" ;
time:calendar = "noleap" ;
|
ORCHIDEE22 : S1(S2) vs S3
ORCHIDEE3 S3 vs S4
