pygetm.momentum module

class pygetm.momentum.CoriolisScheme(*values)[source]

Bases: IntEnum

DEFAULT = 1

ESPELID = 2: Espelid et al. (2000)

OFF = 0

class pygetm.momentum.Momentum(Am: float = 0.0, An: float = 0.0, cnpar: float = 1.0, advection_scheme: AdvectionScheme | None = None, advection_split_2d: AdvectionSplit = AdvectionSplit.FULL, coriolis_scheme: CoriolisScheme = CoriolisScheme.DEFAULT, avmmol: float = 1.8e-06)[source]

Bases: object

Create momentum handler

Parameters:

Am – horizontal viscosity (m2 s-1) If provided, this must be a constant.
An – horizontal diffusivity (m2 s-1) If provided, this must be a constant. It can subsequently be changed through attribute An, which also allows spatially varying diffusivities to be set.
cnpar – parameter for the Crank–Nicolson vertical diffusion solver
advection_scheme – advection scheme
advection_split_2d – directional splitting for advection solver
coriolis_scheme – interpolation method to use to recontruct velocities for Coriolis terms
avmmol – molecular viscosity (m2 s-1)

An

An_uu

An_uv

An_vu

An_vv

SS

SxA

SxB

SxD

SxF

SyA

SyB

SyD

SyF

U

Ui

V

Vi

advU

advV

advance(timestep: float, split_factor: int, tausx: Array, tausy: Array, dpdx: Array, dpdy: Array, idpdx: Array, idpdy: Array, viscosity: Array)[source]

Update depth-explicit transports (pk, qk) and velocities (uk, vk). This will also update their halos.

Parameters:

timestep – (macro) time step (s)
split_factor – number of microtimesteps per macrotimestep
tausx – surface stress (Pa) in x-direction
tausy – surface stress (Pa) in y-direction
dpdx – surface pressure gradient (dimensionless) in x-direction
dpdy – surface pressure gradient (dimensionless) in y-direction
idpdx – internal pressure gradient (m2 s-2) in x-direction
idpdy – internal pressure gradient (m2 s-2) in y-direction
viscosity – turbulent viscosity (m2 s-1)

advance_3d_transport(tp3d: Array, tp2d: Array, vel3d: Array, dp: Array, cor: Array, adv: Array, diff: Array, idp: Array, taus: Array, rr: Array, viscosity: Array, dev: Array, timestep: float)[source]

Advance 3D transports (layer-integrated velocities in m2 s-1) in x- or y-direction, using all relevant tendencies supplied as arguments. This also updates the halos.

Note that 3D velocities (transports divided by thicknesses) are not updated here and therefore will be out of sync with transports after this function returns.

advance_depth_integrated(timestep: float, tausx: Array, tausy: Array, dpdx: Array, dpdy: Array)[source]

Update depth-integrated transports (U, V) and depth-averaged velocities (u1, v1). This will also update their halos.

Parameters:

timestep – time step (s)
tausx – surface stress (Pa) in x-direction
tausy – surface stress (Pa) in y-direction
dpdx – surface pressure gradient (dimensionless) in x-direction
dpdy – surface pressure gradient (dimensionless) in y-direction

advection_scheme

advection_split_2d

advpk

advqk

apply_bottom_friction

avmmol

bottom_friction(u: Array, v: Array, DU: Array, DV: Array, ru: Array, rv: Array, update_z0b: bool = False)[source]

Calculate bottom friction

Parameters:

u – velocity in x-direction (m s-1) in bottom layer [U grid]
v – velocity in y-direction (m s-1) in bottom layer [V grid]
DU – thickness (m) of bottom layer on the U grid
DV – thickness (m) of bottom layer on the V grid
ru – the square of friction velocity (= stress in Pa divided by density), per actual velocity at the layer center on the U grid
rv – the square of friction velocity (= stress in Pa divided by density), per actual velocity at the layer center on the V grid
update_z0b – whether to iteratively update the hydrodynamic bottom roughness

cnpar

corU

corV

coriolis(U: Array, out: Array, x_direction: bool)[source]

Calculate change in transport due to Coriolis force

Parameters:

U – 2d or 3d transport in x- or y-direction (m2 s-1)
out – array to store change in complementary transport (y or x-direction) due to Coriolis force (m2 s-2)

coriolis_scheme

corpk

corqk

dampU

dampV

diffU

diffV

diffpk

diffqk

diffuse_momentum

ea2

ea4

hU_bot

hV_bot

initialize(logger: Logger, tgrid: Grid, runtype: RunType, default_advection_scheme: AdvectionScheme)[source]

logger

pk

qk

rru

rrv

ru

runtype

rv

shear_frequency(uk: Array, vk: Array, viscosity: Array, out: Array)[source]

Calculate squared shear frequency

Parameters:

uk – depth-explicit velocity in x-direction (m s-1)
vk – depth-explicit velocity in y-direction (m s-1)
viscosity – vertical viscosity (m2 s-1)
out – array to store squared shear frequency (s-2)

start()[source]

taub

u1

u_V

u_bot

uadv

udev

uk

update_depth_integrated_diagnostics(timestep: float, skip_coriolis: bool = False, update_z0b: bool = False)[source]

Update 2D momentum diagnostics, including the Coriolis terms that will drive the next 2D update.

Parameters:

timestep – time step (s) to calculate advection of momentum over
skip_coriolis – flag to indicate that Coriolis terms are already up-to-date and do not need recomputing, for instance, after a recent call to advance_depth_integrated()
update_z0b – whether to iteratively update hydrodynamic bottom roughness

update_diagnostics(timestep: float, viscosity: Array, skip_coriolis: bool = False)[source]

Update 3D momentum diagnostics, including the vertical velocity ww, the slow terms that will drive the 2D updates over the next macrotimestep, and the bottom friction and Coriolis terms that will drive the next 3D update. NB the Coriolis update is already done as part of the momentum update itself, so needed only when starting from a restart.

Parameters:

timestep – time step (s)
viscosity – turbulent viscosity (T grid, layer interfaces, m2 s-1)
skip_coriolis – flag to indicate that Coriolis terms are already up-to-date and do not need recomputing