Asymmetry in idealized case, for cloud resolving modelling

Anything related to the idealized simulations in WRF.

Asymmetry in idealized case, for cloud resolving modelling

Postby costantino » Wed Oct 17, 2012 4:21 am

Hi,

I quite new to WRF and I am working with an idealized simulation (starting from the squall line 2d X example) and I want to produce a very tall cloud up to the tropopause (~ 16 km). I use a tropical temperature and humidity profile, up to 40 km. The main idea is to use WRF as a cloud resolving model to study deep convection, with a dx of 250 m, 1500 X-horizontal levels, 200 vertical level, a time step dt of 3s.

To do that, I need a very huge warm bubble (with an horizontal radius of about 150 km and a small vertical one of about 2 km, centred at 2 km of altitude) and several minutes of simulation (typically more than 100, to allow for production of low level U-wind convergence and then enough vertical velocity w to permit cluod formation). An ordinary warm bubble with radius of 10-20 km did not work out (cloud top remains at 2/3 km).

1/ Did anyone of you do similar simulations and have experience with that? Do you confirm the need of such WB to create someting like deep convection? (i.e. to create a cloud that grows up to the tropopause). I know that a wind shear would help, but my aim is to simulate and analyse such a very symmetryc and "clean" case.

2/ After enough time (about 100-150 mins) cloud forms (with a diameter of about 10 km) but it is very little symmetric (at the beginning) and dramatically asymmetryc after a while (~20-30 mins after cloud formation) :cry: . This seems to be very strange, because I DO NOT use SOUNDING (I use symmetric boundary conditions on y, open conditions on both left and right X boundaries, very large domain, NO U-wind, NO V-wind). What can drive this strong asymmetry? numerical errors ? Coriolis (but I think Coriolis is turned off by default, isn't it?)

In addition, I have:
rk_ord=3, diff_opt=2, km_opt=3, khdif and kvdif=0 (but I think these two diffusion coefficients are useless when using Smagorinky parametrization, isn,' it?), damp_opt=3, dampcoef=0.05, zdamp=5000, non_hydrostatic=true, pert_coriolis=false, mix_full_fields=true, h_mom_adv_order=5, v_mom_adv_order=3

Thanks a lot in advance !

Lorenzo
Last edited by costantino on Tue Nov 13, 2012 6:03 am, edited 1 time in total.
costantino
 
Posts: 4
Joined: Thu Sep 20, 2012 10:14 am

Re: Asymmetry in idealized case, for cloud resolving modelli

Postby AchternDiek » Wed Oct 17, 2012 4:41 am

I have no experience with deep convection studies in WRF.
But in general, I would recommend a smaller time step.
time_step= 6*dx is the usual rule.
I prefer something slightly smaller, like 4*dx or 5*dx.

I am not sure if that will help you with your asymmetric cloud.
But with such a small dx and so many vertical levels, I would go with a time_step of 1.0 to 1.5 seconds.

--AchternDiek
AchternDiek
 
Posts: 147
Joined: Wed Jun 06, 2012 6:10 am
Location: Norway

Re: Asymmetry in idealized case, for cloud resolving modelli

Postby costantino » Fri Oct 19, 2012 10:13 am

Thank you very much for your rapid answer. I tried with
a more appropriate time step of 2 sec, for a dx of 450 m
and less vertical levels (150).

Result: the degree of symmetry is always very low.

I will try with a very higher resolution, in order to better
resolve turbulent processes at smaller scales.
costantino
 
Posts: 4
Joined: Thu Sep 20, 2012 10:14 am

Re: Asymmetry in idealized case, for cloud resolving modelli

Postby costantino » Tue Nov 13, 2012 5:41 am

Actually, I can say that using a larger domain, the loss of symmetry appears more rapidly, as well as decreasing the time_step and increasing the time_step_sound
costantino
 
Posts: 4
Joined: Thu Sep 20, 2012 10:14 am

Re: Asymmetry in idealized case, for cloud resolving modelli

Postby franciscaglastra » Fri Sep 05, 2014 4:38 pm

Hi, Constantino, have you resolved the problem?
I was trying the same problem! I want to study deep convection using the squall line in WRF I have started with a tropical profile, and I was trying to fix the sea surface temperature during all simulation to a warmer case say 306 K, but it seems that the simulation does not work, in my case the warm bubble starts convection but rapidly vanish and it doesn't forms clouds anymore...
Any help or advise would be fine!
Thank you, Francisca.
franciscaglastra
 
Posts: 4
Joined: Wed Jun 19, 2013 3:35 pm


Return to Idealized Simulations

Who is online

Users browsing this forum: No registered users and 4 guests