Same output with different configuration

[janac4]

Hello,
I'm working with WRF model over a region covering Panama. My study consists on analysing the output with 5 different physics configuration in the namelist.input. I tried it 2 months ago and I obtained different results according to what was expected (I analyzed T2 and Q2 variables).
Now I'm trying to do it again with few modifications to the namelist but the output I get from the 5 configurations is almost the same. I attach two of the namelists that are supposed to give diferent results but in the end they barely do it.
From my previous study I only changed lat-lon projection for mercator and also changed bdy_control options in the namelist.input as in the forum they suggested me to use specific options for real cases.
In the past these options where configured as below:

 &bdy_control
 periodic_x                          = .true., 
 symmetric_xs                        = .false.,
 symmetric_xe                        = .false.,
 open_xs                             = .false.,
 open_xe                             = .false.,
 periodic_y                          = .true., 
 symmetric_ys                        = .false.,
 symmetric_ye                        = .false.,
 open_ys                             = .false.,
 open_ye                             = .false.,
 /

In the attached files you can see the actual options. Is there any reason the results are so similar eventhough physics is different?
I'm quite desperate :?

 

[Ming Chen]

Hi Jana,
I am not sure what differences you are expecting from these runs? Even if the physics options are different, we generally expect the results are similar. This is because despite the differences between different physics schemes, they should describe the same phenomena on the same physics/dynamics basis.

 

[janac4]

Hello,
I have attached a picture of the scattering plots I obtained when comparing configurations with each other 2 months ago (fig1). You can see that there where different patterns eventhough all the configurations simulated the same phenomena.
In the second picture (fig2) the scatter plots are from my last runs (it includes comparison with GFS analysis files). Between the two pictures, the differences where:
- FIg1 uses bdy_control options for ideal case and fig2 for real cases as you suggested.
- Fig1 uses lat-lon projection and fig2 mercator.
CONUS differs from the others in fig1 because its namelists options are different (as I said in another post, to make it work I had to change bdy_control options to ideal). Does this justify the result?

 

[Ming Chen]

jana,
When you run WRF with different physics/dynamics options and map projections, the results could be similar but they cannot be same. For some cases the results could be very different. You may verify the results against observations, and determine which options/map projections can give you better results.