Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

nwp_diagnostics variables freezed frames

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

Hello.

I have a problem that looks like code bug.

I'm using nwp_diagnostics = 1 in order to get max hourly diagnostic fields. However, they seems like randomly freezing values for several hour frames at random areas in domain. It is hard to explain, so please look into attached animation. You will easy notice freezed updraft values in northern-west map area frames between 17 and 20 UTC.

I'm now using v4.1 but did noticed same problem way before, at least with all way down to v3.9.1 version and with very different setups than one in this example, so I suspect that this is broader issue not neccessary related to my particular build.

As far as I know, this "freezing" happen for MAXUVV:400-1000 mb and MAXREF:1000 m above ground but could be that also other nwp_diagnostic fields freeze also as well.

Can you investigate please?

I have attached my namelist files. Input was GFS 0.25°. If you want to try to replicate exact situation, note that for this particular run, GFS 18z from 2019-05-09 was used and f06h forecast was used as "analyse" to initialize WRF.

TIA,
Ivan
 

Attachments

  • namelist.wps
    970 bytes · Views: 55
  • namelist.input.txt
    7.5 KB · Views: 57
  • updraft.gif
    updraft.gif
    1 MB · Views: 892
I looked through the codes but didn't find anything incorrect. Can you try different cases and check whether the similar "freezing" phenomenon still exists? You may plot time series of some variables for further examination.
 
RCarpenter said:
You can't use adaptive time stepping with nwp_diagnostics.

Hello, thank you for information, I guessed that might be the case, because it looks like registers are not emptied sometimes at full hour mark (for example if because of adaptive time step exact time of full hour is not reached. However, I hoped that by using adjust_output_times = .true. and step_to_output_time = .true. timestep will hit full hour every time so that emptying will happen properly. Does this has sense or problem is somewhere else?

Is this incompatibility between nwp_diagnostics and adaptive time step documented somewhere? I can't find it.
 
Top