Verifying model change was realized

[jzweifel]

Hello CESM-Forums team!

I am posting with a quick question that I'm failing to answer myself.

I've made my first successful change in my user_nl_cam by adding my own altered aerosol emission file as seen below:

! Users should add all user specific namelist changes below in the form of
! namelist_var = new_namelist_value

&chem_inparm
srf_emis_specifier = 'num_a2 -> /glade/work/jzweifel/cases/control_F2000/aerosol_blob_dated.nc'
/
I now want to verify that the change I made was realized in my model results. To do this I thought to compare two history files; one from before the change in my control, and one from after implementing this change. For reference, these were the two files I thought to compare:

experiment: control_F2000.cam.h0.0051-01.nc

control_history: control_F2000.cam.h0.0050-12.nc

What would be the best method to do this? I first tried to use ncview and just graphically analyze the changes, but that seemed a little difficult (but maybe it is just the best way to do this). My advisor suggested I use ncdiff, but that doesn't look like its a module available to download on my account on derecho.

Any thoughts on how to best go about this? If so, let me know how best to help you help me!

Thanks,

Jack Zweifel

 

[jzweifel]

Hello Forum people!

I discovered 'ncdiff' is in the 'nco' module, so after having downloaded that I can now use it!

Now that I want to use it though I have another question.

It occurs to me I should use some formant like this:

ncdiff -v variable_name file1.nc file2.nc diff.nc

My question then is, what variable name should I use? This has been a bit of a tricky thing for me to figure out because the emission variable names are not all one word or joined by underscores. Rather I can see in my cam.input_data_list that my change to my model points like this:

srf_emis_specifier for num_a2 = .......

I tried to put the variable_name to srf_emis_specifier as seen above and was getting an error about too many arguments being present (i'm guessing because the variable name is not conjoined)

Any ideas of how to go forward would be appreciated!

 

[nusbaume]

Just letting anyone who comes across this post know that it appears this question has been answered here:

verifying a change using ncdiff

Hello Forum people! I'm trying to use ncdiff to verify if change that I successfully implemented in my user_nl_cam and ran was actually realized. I'm thinking I should use some formant like this where variable_name is the variable I edited, file1.nc is the last history file before implementing...

bb.cgd.ucar.edu

Have a great day!

Jesse