ERROR with ozone file

[huihuojia@163_com]

I do a hybrid CESM run and write these clauses into the user_nl_cam file.

prescribed_ozone_file = '/R3/cesm_inputdata/inputdata/atm/cam/ozone/ozone_1.9x2.5_L26_1850-2015_rcp45_c101108.nc'

And I got these error message:

/R3/cesm_inputdata/inputdata/atm/cam/ozone/ozone_1.9x2.5_L26_1850-2015_rcp45_c101108.nc 1441792
pio_support::pio_die:: myrank= -1 : ERROR: nf_mod.F90: 671 :
NetCDF: Variable not found
abort:

1. The refcase for this hybrid run is running with nlev=30, and the prescribed_ozone_file is nlev=26, will this cause the problem?
2. The case I run is in 2003, and I check the prescribed_ozone_file. and find time 2003 is not included in this file, will this cause the problem?

 

[eaton]

The prescribed_ozone module does spatial and temporal interpolation, so neither 1. nor 2. are problems.To successfully use this dataset for year 2003 the value of the namelist variable prescribed_ozone_type will need to be 'INTERP_MISSING_MONTHS' and prescribed_ozone_name will need to be 'O3'. 

 

[huihuojia@163_com]

Thanks for your help.I output the pio error, and it is the problem with the ozone variable name, because the default name is ozone not O3.And I have a question, what is the time step for the ozone calculation suppose the dt=1800 for cam? and using prescibed ozone file means that the ozone will keep constant within a month during calculation or it will be interpolated to the time step?Thanks

 

[eaton]

The data is interpolated at each timestep.  But it's not a straightforward linear interpolation because the dataset doesn't contain time samples at each month.  So, for example, to interpolate to 20030101 at 0Z the algorithm starts by recognizing that there are no monthly averages for 20021215 or 20030115.  Those monthly averages are obtained by interpolating linearly between monthly avgs that are present in the dataset.  For the December avg it would use 20001215 and 20051215, and for the January avg it would use 20000115 and 20050115.  Once it has produced data for 20021215 and 20030115 it would interpolate linearly to 20030101 0Z.  The linear interpolation will continue at each timestep using the monthly avgs just constructed until reaching 20030115.  To go past that point a monthly avg at 20030215 will be required, and it will be constructed in the manner just described.