I am very new in this area and wondering if anyone can answer my following questions:
I am running WRF (4.0.3, ideal) to study rounding and truncation. I use the NCL (6.5.0) print command to access variable values in the WRF output file.
I have noticed that some variables have greater precision than others. Is there a way to change the level of precision for individual variables?
I have been using em_b_wave and ideal.exe to initialize the WRF model. The values and variables in the output file are the same no matter how many MPI tasks I use. Why?
Is WRF output exactly the same when using a single MPI task vs. using more than one MPI task? I assume there might be slight differences in rounding and truncation when using different number of processors.
And at the end, I am not sure if the way that I am reading the values and variables from the output files which are binary are correct or not. I am using some NCL scripts now, but have no idea if that is the right way to read the binary output files ...
I would appreciate any help.
Thank you, Sara