MAD-WRF: Setup of passive and active modes

[GPLAND]

Hi all,

I'm trying to setup the two MAD-WRF (4.3v) configurations referred in Jimenez et al. 2022: 1) Passive mode without microphysics parameterization, 2) Active mode with Thomson and a nudging of hydrometeors during the first hour.

I'm not sure, but I understand that mp_physics = 96, is a way of saying that MAD-WRF is selected with no microphysics applied. This assumption is compatible with the statment in the README.madwrf for passive mode ("the user should set mp_physics = 96 and use_mp_re = 0")

Additionally, in principle the active mode should contain Thompson microphysics active.

Therefore, the namelist of both options should contain:

1) Passive:

mp_physics = 96,
madwrf_opt = 1,
madwrf_cldinit = 1,
madwrf_dt_relax = 60.0,
madwrf_dt_nudge = 60.0,
use_mp_re = 0,

2) Active:

mp_physics = 8,
madwrf_opt = 2,
madwrf_cldinit = 1,
madwrf_dt_relax = 60.0,
madwrf_dt_nudge = 60.0,
use_mp_re = 0,

However, the option 2 is not stable and a Segmentation Fault is found always at he beginning (It is not the case when mp_physics is changed to 96).

Anybody knows if the setup of active and passive modes I selected is correct?. If it is the case, any possible explication to the Segmentation Fault found with option 2?

Thanks in advance.

G.P. Landa

 

[jimenez]

Hello,

Your configuration looks correct. It is as described in the WRF/doc/README.madwrf file with the MAD-WRF documentation. You can try compiling WRF in debugging mode (compile -d or compile -D) to obtain more information about the segmentation fault.

If you are ingesting observations to improve the cloud initialization, you need to add entries to the Metgrid table in WPS before running Metgrid. This will be fixed in upcoming WPS versions. For the moment you can look at the pull request ("files changed" tab):

Adding entries to the Metgrid table for MAD-WRF by pedro-jm · Pull Request #206 · wrf-model/WPS

MAD-WRF requires the following entries in the metgrid table. The variables are optional and can be used to improve the cloud analysis in Real.

github.com

But this may not resolve your segmentation fault problem.

Pedro.

 

[GPLAND]

Thanks for your reply Pedro.

I'll try the debugging mode and see what I found.

Regarding the ingest of observations, Metgrid table has been changed and according to what is showed in the outputs, it seems that satellite data are ingested correctly.

Thanks.

G. P. Landa

 

[guolc_IUM]

Hi all,

I'm trying to setup the two MAD-WRF (4.3v) configurations referred in Jimenez et al. 2022: 1) Passive mode without microphysics parameterization, 2) Active mode with Thomson and a nudging of hydrometeors during the first hour.

I'm not sure, but I understand that mp_physics = 96, is a way of saying that MAD-WRF is selected with no microphysics applied. This assumption is compatible with the statment in the README.madwrf for passive mode ("the user should set mp_physics = 96 and use_mp_re = 0")

Additionally, in principle the active mode should contain Thompson microphysics active.

Therefore, the namelist of both options should contain:

1) Passive:

mp_physics = 96,
madwrf_opt = 1,
madwrf_cldinit = 1,
madwrf_dt_relax = 60.0,
madwrf_dt_nudge = 60.0,
use_mp_re = 0,

2) Active:

mp_physics = 8,
madwrf_opt = 2,
madwrf_cldinit = 1,
madwrf_dt_relax = 60.0,
madwrf_dt_nudge = 60.0,
use_mp_re = 0,

However, the option 2 is not stable and a Segmentation Fault is found always at he beginning (It is not the case when mp_physics is changed to 96).

Anybody knows if the setup of active and passive modes I selected is correct?. If it is the case, any possible explication to the Segmentation Fault found with option 2?

Thanks in advance.

G.P. Landa

1、are madwrf_dt_relax = 60.0, and madwrf_dt_nudge = 60.0 need in passive mode?

2、I still can't understand the options of madwrf_dt_relax and madwrf_dt_nudge. Can someone help me explain this again? Thanks !!

 

[guolc_IUM]

1、are madwrf_dt_relax = 60.0, and madwrf_dt_nudge = 60.0 need in passive mode?
View attachment 15411
2、I still can't understand the options of madwrf_dt_relax and madwrf_dt_nudge. Can someone help me explain this again? Thanks !!

I have run a case with madwrf_dt_relax and madwrf_dt_nudge in passive mode. wrfout does not contain tr_qcloud and tr_qice. so I think madwrf_dt_relax and madwrf_dt_nudge are not necessary in passive mode.