How to output temperature tendency terms

[cyh082]

Hello,

I would like to output temperature tendency terms.
I do not have much knowledge on WRF physics, dynamics, and code structure.
Could you answer my following questions?

1) How is total temperature tendency determined in WRF model?
Total tendency for temperature may be sum of tendencies from horizontal/vertical advection, microphysics, cumulus, PBL, radiation schemes, and so on.
Where can I check the above?

2) How can I output all temperature tendency terms?
I found that there are some entries in Registry.EM_COMMON, of which description mentions "theta tendency". For example, h_diabatic, RTHSHTEN, RTHCUTEN, RTHFTEN, RTHRATEN, RTHBLTEN.
Can I output tendency terms by modifying Registry.EM_COMMON file?

Thank you so much.

Yonghan

 

[Ming Chen]

Hi Yonghan,
To answer your question,
(1) the diabatic heating terms include the following:
"h_diabatic": this is from microphsics scheme
"RTHCUTEN": this is from cumulus scheme
"RTHSHTEN": this is from shallow convection
"RTHRATEN": this is from radaition
"RTHBLTEN": this is from boundary layer
The sum of the above terms is the total diabatic heating.

(2) To output these variables, please look at Registry/Registry.EM_COMMON, in which we can find all the above variables. By default they are not included in wrfout files. We can simply add 'h' to the 8th column in the line that defines the variable. For example, below is the original line that defines 'h_diabatic':
state real h_diabatic ikj misc 1 - rdu "h_diabatic" "MICROPHYSICS LATENT HEATING" "K s-1"
We can change it to:
state real h_diabatic ikj misc 1 - rhdu "h_diabatic" "MICROPHYSICS LATENT HEATING" "K s-1"
(Please make the same modification for all other variables (RTHCUTEN, RTHSHTEN, etc.)

(3) Then you need to type ./clean -a and recompile WRF. With the above changes, these variables will be included in wrfout files

 

[cyh082]

Hi Ming,

First of all, thank you so much for your kind and detailed explanation!

I successfully obtained each temperature tendency term based on the method you suggested.

I have two additional questions now.

I got WRF outputs with one-hour interval (with namelist variable, frames_per_outfile=1).
Each tendency term has four dimensions (Time, bottom_up, south_north, west_east), and "Time" dimension size is one.
And each term has unit of K s-1.

1) I used time step of 120 seconds.
In my case, temperature tendency variable in WRF output (for example, at 0000 UTC 20 April 2022) is temperature tendency valid at 0000 UTC 20 April 2022, and temperature value for next time (after 120 seconds) is computed by adding this tendency times 120 to temperature value at this time (0000 UTC 20 April 2022). Am I right?

2) If I would like to obtain averaged temperature tendency for one hour, then should I make WRF output for every time step (i.e., 120 sec)?
I think it needs lots of disk space.
Is there any way to output hourly-averaged temperature tendency by modifying Registry?

Thanks!

Yonghan

 

[Ming Chen]

Hi Yonghan,
Please see my answers below:
1) I used time step of 120 seconds.
In my case, temperature tendency variable in WRF output (for example, at 0000 UTC 20 April 2022) is temperature tendency valid at 0000 UTC 20 April 2022, and temperature value for next time (after 120 seconds) is computed by adding this tendency times 120 to temperature value at this time (0000 UTC 20 April 2022). Am I right?

All wrfout variables are instantaneous. So yes the temperature tendency variable in WRF output (for example, at 0000 UTC 20 April 2022) is temperature tendency valid at 0000 UTC 20 April 2022. The tendency at next time step is recalculated based on the situation at that time step. Note that tendency is not accumulative.

2) If I would like to obtain averaged temperature tendency for one hour, then should I make WRF output for every time step (i.e., 120 sec)?
This sounds not feasible.

Is there any way to output hourly-averaged temperature tendency by modifying Registry?
In WRFV4.4 (we will release this version very soon. Please pay attention to wrf-news announcement), we allow for output of accumulative tendencies. I suppose this new capability will meet your need.

 

[cyh082]

Hi Ming,

Thank you for your kind answers!

That's a great news!

I appreciate your kind help.

Yonghan

 

[cyh082]

Hi Ming,

I just compiled WRF version 4.4 and tested a new capability of "accumulated physics tendency".

I think I worked! Thanks!

I have two questions.

1) Now, accumulated physics tendency variable has unit of K.
If I set output interval of one hour, then does accumulated tendency mean "accumulated for one hour"?

2) There are a total of 6 variables for theta.
Isn't there accumulated tendency from horizontal/vertical advection?

Thank you for your kind help.

Yonghan