Section: ABLMesoForcing

This section is for setting mesoscale forcing parameters for mesoscale-to-microscale coupling (MMC). With the exception of ABL.mesoscale_forcing and ABL.tendency_forcing, the parameters below apply to both ABLMesoForcingMom and ABLMesoForcingTemp for momentum and temperature, respectively.

ABL.mesoscale_forcing

type: String, optional

Path to a netcdf file with mesoscale data. This datafile should have dimensions “ntime” and “nheight”. Time-height-varying variables include “momentum_u”, “momentum_v”, and “temperature” (virtual potential temperature); time-varying variables include “tflux” (kinematic heat flux) and “transition_height” (optional, only used if ABLMesoForcing*.forcing_transition is specified without a ABLMesoForcing*.constant_transition_height). The time-height data are stored as flattened 1-D arrays, wherein the height dimension changes the fastest.

ABL.tendency_forcing

type: Boolean, optional, default = false

If “true”, momentum and temperature source terms are directly provided within the ABL.mesoscale_forcing netcdf file; ABLMesoForcing*.forcing_scheme is ignored. Typically, these source terms are determined from mesoscale model tendencies.

If “false”, the input velocity and temperature time-height data are used for profile assimilation forcing schemes.

ABLMesoForcing*.forcing_scheme

type: String, mandatory

Options include: “direct” or “indirect” profile assimilation. Direct profile assimilation (DPA) is a strong enforcement of the input mesoscale profiles. Height-varying source terms are calculated at every time step to drive the error between the instantaneous planar-averaged fields and the mesoscale data to zero. This strategy implies a high level of confidence in the specified mesoscale conditions and can result in excessive turbulence resolved in the microscale.

In comparison, indirect profile assimilation weakly enforces the mesoscale profiles by using an approximation to the DPA forcing. This is an engineering approach that allows for local deviation from the mesoscale mean quantities and allows the microscale flow to find its own equilibrium between mean profiles and the associated resolved turbulence. The approximate forcing is derived by applying a polynomial regression to the DPA forcing profile.

ABLMesoForcing*.control_gain

type: Real, optional, default = 0.2

A relaxation factor used to scale the magnitude of the source term, which may be interpreted as an inverse time scale with units of 1/s.

ABLMesoForcing*.forcing_transition

type: String, optional, default = “none”

Used to specify two forcing schemes for upper and lower regions under more complicated MMC modeling scenarios, e.g., with complex background conditions and/or uncertainty in the mesoscale data. The default is apply the ABLMesoForcing*.forcing_scheme for the full mesoscale profile over the entire microscale domain. Other options include “directToConstant”, “indirectToConstant”, and “indirectToDirect”.

When blending to a constant forcing profile, the slope of the forcing profile at the transition height (specified by ABLMesoForcing*.constant_transition_height or provided in the ABL.mesoscale_forcing datafile) is linearly reduced to 0 from the transition height up to the transition height + ABLMesoForcing*.transition_thickness.

When blending to the DPA forcing profile (e.g., in the free atmosphere), the two forcing profiles are linearly blended from one to the other over the ABLMesoForcing*.transition_thickness, starting from the transition height as described above.

Indirect Profile Assimilation

The following parameters are specific to the IPA scheme (ABLMesoForcing*.forcing_scheme = “indirect”). At the moment, only third-order polynomial regression is supported.

ABLMesoForcing*.weighting_heights

type: List of Reals (has to be same length as ABLMesoForcing*.weighting_values), optional

Height(s) in meters at which IPA regression weights are provided.

ABLMesoForcing*.weighting_values

type: List of Reals (has to be same length as ABLMesoForcing*.weighting_heights), optional

IPA regression weights at the corresponding ABLMesoForcing*.weighting_heights. The default behavior is to use uniform weighting. Nonuniform weighting is generally ill-advised as runaway positive or negative forcing values may be possible.

ABLMesoForcing*.normalize_by_zmax

type: Boolean, optional, default = false

If “true”, the height coordinate is normalized by the domain height when performing the IPA regression. Provided for consistency with a legacy solver implementation to improve conditioning of the regression matrix but should not be needed.

Partial Profile Assimilation

The following parameters are for “partial” profile assimilation, enabled by ABLMesoForcing*.forcing_transition being not set to “none”. This will only partially apply the instantaneous IPA forcing profiles over the simulation domain. Above a specified transition layer, a secondary forcing profiles may be applied.

ABLMesoForcing*.transition_thickness

type: Real

The thickness of the layer over which the forcing scheme transitions from the lower scheme to the upper scheme.

ABLMesoForcing*.constant_transition_height

type: Real

The base of the transition layer, which is invariant for the duration of the entirety of the simulation. To specify a time-varying transition layer height that, e.g., tracks the evolution of the ABL height, omit this parameter and include the time-varying transition_height variable within the ABL.mesoscale_forcing datafile.