Why are model results reset every 24 hours?

[rjsdn613]

I am conducting an experiment from October 1, 2016 to October 6, 2016.

I have the following met_em files:
met_em.d01.2016-10-01_00:00:00.nc
met_em.d01.2016-10-02_00:00:00.nc
met_em.d01.2016-10-03_00:00:00.nc
met_em.d01.2016-10-04_00:00:00.nc
met_em.d01.2016-10-05_00:00:00.nc
met_em.d01.2016-10-06_00:00:00.nc

met_em.d02.2016-10-01_00:00:00.nc
met_em.d02.2016-10-02_00:00:00.nc
met_em.d02.2016-10-03_00:00:00.nc
met_em.d02.2016-10-04_00:00:00.nc
met_em.d02.2016-10-05_00:00:00.nc
met_em.d02.2016-10-06_00:00:00.nc

I have enhanced the u200 in a specific region of all the met_em files. I then ran real.exe and wrf.exe.
However, the results of the model experiment show that the u200 enhancement is only maintained at 00:00 on October 1st, and it is not maintained after 01:00 on October 1st. I also enhanced the u200 in the met_em file for October 2nd, but the model experiment shows that the u200 is not enhanced again at 00:00 on October 2nd.
As a result, the u200 is only enhanced at 00:00 on October 1st, and it is not properly enhanced at all times after that.
Why is that?



shoud i have to run restart option in this case?

or use io_form_auxinput2 = 2 ????

 

[icastorm]

Hi, I think there are a couple issues here.

First, as I understand it, the met_em files are only used to generate the initial conditions and the boundary conditions. Therefore, if the enhanced winds are located outside of the influence of the boundary conditions, then any changes to the met_em files will be lost after the initial time as the model integrates forward. In other words, the model registers a change to the initial conditions, but all other forcing from the met_em files is done at the boundaries from then on.

That is the second problem: Because WRF is a full-physics model, changes to the initial conditions that aren't reflected by the rest of the conditions in the simulation (in this case, stronger winds than the atmosphere would otherwise produce) would quickly be absorbed into the flow and not maintained. If you plot the wind fields hourly, you will see this happen.

As far as I know, there is no way to maintain an enhanced region of wind as you describe without altering WRF itself, unfortunately. You could modify your domains so that the wind enhancement is near or on the boundary of the outermost domain, although that wouldn't be exactly what you're trying to do here, and the impact would be further dampened by having the wind perturbation downstream of the final domain.

 

[george_victor]

Hi!
I see that you've modified met files in a layer not very close to the boundary and might most probably be the issue with your simulation, as rightfully pointed out by @icastorm. To confirm if this is the reason, you may check the wrfinput files generated from real.exe execution and make sure that the changes you made are visible there. If the changes are visible in wrfinput files and not in the simulation, what @icastorm said should be correct.