Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Many thanks to Dave Bailey (NCAR) for getting this ball rolling.
The history module has been generalized to allow output at different
frequencies. Five output frequencies ('1', 'h', 'd', 'm', 'y') are
available simultaneously during a run. The same variable can be
output at different frequencies (say daily and monthly) via its
f_<variable> string in namelist.
Added max_nstrm=5 to ice_domain_size.F90, one for each output frequency
available, and made the namelist parameters histfreq and histfreq_n
into corresponding arrays. The namelist flags for history variables
have been changed from logicals (true/false) to character strings
corresponding to histfreq or 'x' for none (but note that grid variable
flags are still logicals).
For example, in namelist
histfreq='1', 'h', 'd', 'm', 'y'
histfreq_n=1,6,0,1,1
f_hi = '1'
f_hs = 'h'
f_Tsfc = 'd'
f_aice = 'm'
f_meltb = 'mh'
f_iage = 'x'
Here, hi will be written to a file on every timestep, hs will be
written once every 6 hours, aice once a month, meltb once a month AND
once every 6 hours, and Tsfc and iage will not be written. If there
are no namelist flags with a given histfreq value, or if an element of
histfreq_n is 0, then no file will be written at that frequency.
The output period can be discerned from the filenames as in the
released code (this has not changed).
From an efficiency standpoint, it is best to set unused frequencies in
histfreq to 'x'. Having output at all 5 frequencies takes nearly 5
times as long as for a single frequency. If you only want monthly
output, the most efficient setting is histfreq='m','x','x','x','x'. The
code counts the number of desired streams (nstreams) based on histfreq.
The history variable names must be unique for netcdf, so in cases where a
variable is written at more than one frequency, the variable name is
appended with the frequency in files after the first one. In the example
above, meltb is called meltb in the monthly file (for backward
compatibility with the default configuration) and meltb_h in the 6-hourly
file.
There are still 9 history variables hard-coded for instantaneous output,
as before. They are output at the frequency given by their namelist flag.
The names of two category history variables had to be changed from aice,
vice to aicen, vicen to make the roots unique. They are now consistent
with the other category history variables.
Notes:
Using the same frequency twice in histfreq will have unexpected
consequences and currently will cause the code to abort. It is not
possible at the moment to output averages once a month and also once
every 3 months, for example.
CCSM's special units are not included in this version.
The units for Tair were listed as degrees C, but the field in the (old)
file was in K. Now they are both C.
Grid variables are currently written to all files. I'd prefer to leave
the minimum information in the history files, and maybe write the rest in
the initial condition file, to cut down on redundancy. Suggestions or
opinions welcome!
I did not test every possible parameter combination; do let me know if
you find one that doesn't work correctly or as expected.- Loading branch information
1 parent
d607f8c
commit ab1c53e