Skip to content

Interior nudging

In order to improve accuracy of the LAM model, it is possible to nudge the interior of the domain as well. The whole configuration is simple, and it requires only small changes in the DYNAMICO run_dynamico.def file. But first, we need to understand the nudging in the context of the LAM.

Nudging in LAM

In DYNAMICO, the nudging can happen on temperature, wind, surface pressure, and/or humidity. There is a set of four parameters guide_* to turn the nudging on/off:

Variable DYNAMICO parameter Global default value LAM default value
Temperature guide_T .FALSE. .TRUE.
Wind guide_U .FALSE. .TRUE.
Surface pressure guide_PS .FALSE. .TRUE.
Humidity guide_Q .FALSE. .TRUE.

Furthermore, nudging can be applied to the outside and the inside of the domain. The relaxation time in seconds is set with parameters:

  • <VAR>_relax_out (outside nudging with default value dt)
  • <VAR>_relax_in (interior nudging with default value 0)

Where <VAR> is one of: T, U, PS, Q. The timestep dt is then divided by <VAR>_relax_out and/or <VAR>_relax_in to produce the coefficients used when applying nudging.

Default nudging requires all variables

LAM model is based on the nudging_zone=area parameter which forces the use of all four variables for the nudging on the border around its domain (lateral boundary conditions). If we don't want to use a specific variable for the interior nudging, we cannot simply set guide_*=.FALSE..

Applying interior nudging

Even though, all variables are used for nudging in LAM, they are not used for the interior nudging by default, because their relaxation coefficient is 0. In order to turn on interior nudging, set the <VAR>_relax_in to a positive value (12 hours, i.e. 43000 seconds, is a common choice):

U_relax_in=43000 #(1)!
T_relax_in=43000 #(2)!
  1. Apply nudging inside the LAM domain to wind variable.
  2. Apply nudging inside the LAM domain to temperature variable.

Here, only the wind and temperature will be nudged in the interior of the domain.

Interior nudging relaxation

Because the initial value of the model is quite different from nudging data, it might be necessary to use a quite big value, e.g. 43000, so that the nudging input values are not applied to much to the simulation and they don't produce garbage. Remember, this value will divide the dt to create the final nudging coefficient!

Changing vertical application of nudging

The nudging is applied uniformly in the vertical direction. However, we often want to apply nudging only to the upper part of the atmosphere. In order to do so, we can set up the following parameters:

nudging_vert_profile_type = "tanh" #(1)!
nudging_vert_profile_press_lev = 70000 #(2)!
  1. Use hyperbolic tangent (tanh) vertical profile to apply the nudging at changing rate.
  2. Vertical nudging from above 700 hPa.