MYNN PBL vs YSU PBL convection

[wallis]

I am comparing MYNN vs. YSU as a PBL for 3km forecasts in Scandinavia.
Namelist looks like this:
mp_physics = 6, 6, 6,
ra_lw_physics = 4, 4, 4,
ra_sw_physics = 4, 4, 1,
radt = 3, 3, 3,
sf_sfclay_physics = 5, 5, 5,
sf_surface_physics = 4, 4, 4,
bl_pbl_physics = 5, 5, 5,
bl_mynn_mixlength = 2,
bldt = 0, 0, 0,
cu_physics = 0, 0, 0,
cudt = 5, 5, 5,

MYNN matches observations much closer, but it very frequently predicts no thunderstorms when in reality and with YSU PBL we see strong thunderstorms forming.

What can be done to improve this?

 

[Pechudin]

I'm no expert but AFAIK YSU is more geared towards convective conditions, which would be your use-case. MYNN is better for more complex terrain, though, especially downslope windstorms.

If you are willing to go to sub-1km resolution, there is a gray-zone scheme (Shin-Hong, which reduces to YSU at large scale) which is even better for convective boundary layers (bl_pbl=11, I think).

I am not sure on microphysics though.

 

[wallis]

Unfortunately we don't have the resources to go to such high resolution. Can we use a cu_physics option that might help?

 

[wallis]

I understand HRRR is using MYNN and I think it resolves convective storms quite well?

 

[Pechudin]

Unfortunately we don't have the resources to go to such high resolution. Can we use a cu_physics option that might help?

Unsure, sorry. cu_physics is usually used on > 3km resolution, but it could not hurt to give it a whirl. I am also relatively new, so take my advice with a grain of salt.

 

[Pechudin]

Also is the problem not really the "resolving" of clouds, but the initiation, i.e. triggering of convection? In that case PBL is important. Is your problems the plumes not starting or rising high enough?

 

[haiqingsong]

I am comparing MYNN vs. YSU as a PBL for 3km forecasts in Scandinavia.
Namelist looks like this:
mp_physics = 6, 6, 6,
ra_lw_physics = 4, 4, 4,
ra_sw_physics = 4, 4, 1,
radt = 3, 3, 3,
sf_sfclay_physics = 5, 5, 5,
sf_surface_physics = 4, 4, 4,
bl_pbl_physics = 5, 5, 5,
bl_mynn_mixlength = 2,
bldt = 0, 0, 0,
cu_physics = 0, 0, 0,
cudt = 5, 5, 5,

MYNN matches observations much closer, but it very frequently predicts no thunderstorms when in reality and with YSU PBL we see strong thunderstorms forming.

What can be done to improve this?

Yes，I have the same problem. So，What is your solution？

 

[haiqingsong]

I am comparing MYNN vs. YSU as a PBL for 3km forecasts in Scandinavia.
Namelist looks like this:
mp_physics = 6, 6, 6,
ra_lw_physics = 4, 4, 4,
ra_sw_physics = 4, 4, 1,
radt = 3, 3, 3,
sf_sfclay_physics = 5, 5, 5,
sf_surface_physics = 4, 4, 4,
bl_pbl_physics = 5, 5, 5,
bl_mynn_mixlength = 2,
bldt = 0, 0, 0,
cu_physics = 0, 0, 0,
cudt = 5, 5, 5,

MYNN matches observations much closer, but it very frequently predicts no thunderstorms when in reality and with YSU PBL we see strong thunderstorms forming.

What can be done to improve this?

MYNN matches observations much closer.
Which variable did you concern ？

 

[wallis]

Also is the problem not really the "resolving" of clouds, but the initiation, i.e. triggering of convection? In that case PBL is important. Is your problems the plumes not starting or rising high enough?

Convection is eventually triggered, but much later in the day and weaker than observations. YSU correctly predicts it.

MYNN 2m temp is cooler by 1-2c, perhaps this is the reason? Why would a different PBL scheme make so much difference?

 

[Pechudin]

Convection is eventually triggered, but much later in the day and weaker than observations. YSU correctly predicts it.

MYNN 2m temp is cooler by 1-2c, perhaps this is the reason? Why would a different PBL scheme make so much difference?

Well, convection is driven by surface heating and subsequent transport of heat and moisture in the free atmosphere above. If this transport is underestimated, then there will not be enough energy for the thermals to rise high enough above the condensation level to trigger condensation and latent heat release.

Therefore it stands to reason that PBL representation is very important for convection initiation.