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

FARMS time series

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.


New member

I was wondering if anyone else in the community has experienced issues with the Fast All-Sky Radiation Model (FARMS), specifically, the influence that its timestep seems to exert on plots of the time series for surface variables.

The time series that are written on the *.TS file, show a random stepping behavior at certain positions (circled in the example image included below for the V-wind component). I have noticed that the frequency of those step matches radt (the minutes between two successive radiation physics calls). Decreasing radt does not eliminate the appearance of those steps completely however.

For reference, I am including below a list of the physics models I am currently using :
 mp_physics                          = 8,       8,
 ra_lw_physics 			     = 4,       4,
 ra_sw_physics 			     = 4,       4,
 radt 				     = 4,       4,
 sf_sfclay_physics		     = 1,       1,
 sf_surface_physics 		     = 2,       2,
 bl_pbl_physics 		     = 5,       5,
 bldt 				     = 0,       0,
 bl_mynn_tkeadvect 		     = .true.,  .true.,
 cu_physics 			     = 0,       0,
 cu_rad_feedback 		     = .false.,  .false.,
 cudt 				     = 0,       0,
 bl_mynn_edmf 			     = 0,       0,
 shcu_physics 			     = 0,       0,
 isfflx 			     = 1,
 ifsnow 			     = 1,
 icloud 			     = 1,
 icloud_bl 			     = 0,
 num_soil_layers 		     = 4,
 sf_urban_physics 		     = 0,       0,
 num_land_cat 			     = 21,
 aer_opt 			     = 1,
 swint_opt 			     = 2,
 usemonalb 			     = .true.,



  • Timeseries.png
    24.5 KB · Views: 731
Just for some additional info: this stepping behavior appears only when plotting the timeseries from the *.TS file. The timeseries taken from the model level files (*.UU, *.VV, *.WW, etc.) appear to be clean on the other hand, even those that correspond to the lowest model level (near the surface).
I have a few questions:
(1) If you don't activate FARMS, how about the time series at the same time and location?
(2) Is this feature specific for V10, or it is common for all the surface variables?
(3) It seems to appear randomly, not always corresponds to the time step for radiation. Is this correct?
(4) Which version of WRF did you run for this case?
Thanks for getting back to me.

(1) I am currently running additional simulations with different radiative transfer models. I will let you know if theissue persists once I have more data.
(2) It is common for other variables as well. It is also prominent for T2 for example.
(3) I am not sure if it appears randomly. I thought that the steps corresponded to successive calls to the radiative transfer model. Their frequency of appearance seems to be steady. They do however show up only in certain parts of the time-series, where there seems to be a steeper gradient for GHI and/or the plotted variable.
(4) This is from the latest version of WRF (4.2.1)

I was expecting that plots of the irradiance variables would exhibit these temporal discontinuities, since calls to FARMS are more infrequent than model steps, as described by Jimenez et al (2016), but I am concerned that they also appear on other variables as well. It seems that there is some sort of influence.
Please take a look at these code changes and let me know whether they fixed your problem.
A quick update:
I've checked the code changes now that they have been merged into the WRF v4.2.2 branch. I don't think they change the behavior of the surface timeseries plots. On the other hand, I have found that enabling DFI and/or setting swint_opt = 1 attenuates this issue, but doesn't eliminate it completely. This is not exclusive to FARMS (swint_opt = 2) by the way. I have gotten similar results with the full RRTMG model.
Thanks for the update. I will talk to our physics experts and see whether we can find the possible reasons.
It seems that FARMS output does not affect the simulation. If you simulations with/without FARMS activated the output should be the same. Hence, the discontinuities may be due to radt. Can you run a case that calls radiation every time step, and see whether the discontinuities disappear? Please keep me updated about the results. Thanks.