derive_transport

Attributes

model_trans_dir

weight_folder

monthly_transport

filepath

gc_metpath

month

Functions

interpolate_T(ds_met, var, zb3D, lon, zx2D, latx2D)

Interpolate dataset variable from 3D to 2D grid using xesmf

interpolate_cflux(ds_met, var, zb3D, lon, zx2D, latx2D)

Interpolate dataset variable from 3D to 2D grid using xesmf.

pchip2(x, y, z, xi, yi)

P-chip interpolation on 2-D. x, y, xi, yi should be 1-D

infer_eddy_agrid(dsvqt, dsvt, dswqt, dswt, y2D, ym2D)

This infers the eddy trasport tensor following the approach of Bachmann et al. (2015)

validation_tracer_agrid(dsvqt, dsvt, dswqt, dswt, tn, ...)

Flux-gradient calculation if wishing to validate the eddy-flux derivation.

smoothKzz(Kz[, limit, lower])

Kzz is hard to derive so smooth over unrealistically high diffusion.

correct_eddy(K, zm2D, z2D, ym2D, y2D)

Kyy can't be negative. Put in check/correction as Plumb & Mahlman 1986

derive_ustar(K, Kzy_ygrid, Kyz_zgrid, zm2D, ym2D)

Take anti-symmetric part of eddy transport tensor to derive

wstar_potentialtemp(dsTwt, dsvwt, H)

Calculate w* (eddy advective component) using potential temperature (not used)

make_nondivergent(w_in, v_in, cosc, cose, H, z, zm, y, ...)

Ensure residual wind fields are non-divergence.

add_convection(ds2d)

Add convective parameters straight from MERRA2 met data

make_2D_yearly_files(start_year[, end_year])

Take monthly 2D transport and make CF compliant 3D netcdf files

Module Contents

derive_transport.model_trans_dir
derive_transport.weight_folder = '/user/home/lw13938/work/TwoDmodel/weights'
derive_transport.monthly_transport = '/user/home/lw13938/work/TwoDmodel/TransportParameters'
derive_transport.filepath = '/user/home/lw13938/work/GCClassic.13.3.4/rundirs/Tracer_outputs/'
derive_transport.gc_metpath = '/user/home/lw13938/shared/GEOS_CHEM/data/ExtData/GEOS_4x5/MERRA2/'
derive_transport.interpolate_T(ds_met, var, zb3D, lon, zx2D, latx2D)

Interpolate dataset variable from 3D to 2D grid using xesmf

Parameters:

  • ds_met (dataset) – xarray dataset containing variable to interpolate

  • var (string) – variable to interpolate

  • zb3D (array) – Scale height of 3D model

  • lon (array) – Longitude of 3D model

  • zx2D (array) – Scale height of 2D model

  • latx2D (array) – Latitude of 2D model

Returns:

xarray DataArray of interpolated variable

Return type:

DataArray

derive_transport.interpolate_cflux(ds_met, var, zb3D, lon, zx2D, latx2D)

Interpolate dataset variable from 3D to 2D grid using xesmf. Basically the same as interpolate_T but need some slight modifications.

Parameters:

  • ds_met (dataset) – xarray dataset containing variable to interpolate

  • var (string) – variable to interpolate

  • zb3D (array) – Scale height of 3D model

  • lon (array) – Longitude of 3D model

  • zx2D (array) – Scale height of 2D model

  • latx2D (array) – Latitude of 2D model

Returns:

xarray DataArray of interpolated variable

Return type:

DataArray

derive_transport.pchip2(x, y, z, xi, yi)

P-chip interpolation on 2-D. x, y, xi, yi should be 1-D and z.shape == (len(x), len(y))

Taken from: https://scipy-user.scipy.narkive.com/FG5DVM1l/2-d-data-interpolation This ensures monoticity for interpolation (splines can over/undershoot)

derive_transport.infer_eddy_agrid(dsvqt, dsvt, dswqt, dswt, y2D, ym2D)

This infers the eddy trasport tensor following the approach of Bachmann et al. (2015) Input are data arrays of v and w winds, and dataset of transport tracers all interpolated to the 2D model resolution.

Parameters:

  • dsvqt (dataset) – Concentrations on same grid as v-field

  • dsvt (dataset) – Met data on same grid as v-field

  • dswqt (dataset) – Concentrations on same grid as w-field

  • dswt (dataset) – Met data on same grid as w-field

  • y2D (array) – y-coordinate at grid cell edges

  • ym2D (array) – y-coordinate at grid cell mid-points

Returns:

Lists containing the eddy flux tensor and gradient flux fields

Return type:

list

derive_transport.validation_tracer_agrid(dsvqt, dsvt, dswqt, dswt, tn, y2D, ym2D)

Flux-gradient calculation if wishing to validate the eddy-flux derivation. Note, this is not currently used.

Parameters:

  • dsvqt (dataset) – Concentrations on same grid as v-field

  • dsvt (dataset) – Met data on same grid as v-field

  • dswqt (dataset) – Concentrations on same grid as w-field

  • dswt (dataset) – Met data on same grid as w-field

  • tn (int) – Tracer number used for validation

  • y2D (array) – y-coordinate at grid cell edges

  • ym2D (array) – y-coordinate at grid cell mid-points

Returns:

Gradient flux fields.

Return type:

arrays

derive_transport.smoothKzz(Kz, limit=50.0, lower=False)

Kzz is hard to derive so smooth over unrealistically high diffusion. Kzz is generally higher in the lower atmosphere. Set upper Kzz limit to 50 m2/s (Plumb and Mahlman max is around 20) as upper limit, and reset as the average of surrounding grid squares.

Parameters:

  • Kz (array) – Kzz component of eddy-transport tensor

  • limit (float, optional) – Upper limit of Kzz. Defaults to 50.

  • lower (bool, optional) – True if setting a lower limit. Defaults to False.

Returns:

Smoothed Kzz tensor component.

Return type:

array

derive_transport.correct_eddy(K, zm2D, z2D, ym2D, y2D)

Kyy can’t be negative. Put in check/correction as Plumb & Mahlman 1986

Parameters:

  • K (list) – Eddy-transport tensor

  • zm2D (array) – z-coordinate mid-points in 2D model

  • z2D (array) – z-coordinate edges in 2D model

  • ym2D (array) – y-coordinate mid-points in 2D model

  • y2D (array) – y-coordinate edges in 2D model

Returns:

Corrected Eddy-transport tensor, the Kzy tensor component

interpolated to other edges, diffusivity on different grids

Return type:

list, arrays

derive_transport.derive_ustar(K, Kzy_ygrid, Kyz_zgrid, zm2D, ym2D)

Take anti-symmetric part of eddy transport tensor to derive residual advective transport.

Parameters:

  • K (list) – Eddy-transport tensor.

  • Kzy_ygrid (array) – Kzy component on v-wind grid.

  • Kyz_zgrid (array) – Kzy component on w-wind grid.

  • zm2D (array) – z-coordinate mid-points in 2D model.

  • ym2D (array) – y-coordinate mid-points in 2D model.

Returns:

residual v and w advection.

Return type:

array

derive_transport.wstar_potentialtemp(dsTwt, dsvwt, H)

Calculate w* (eddy advective component) using potential temperature (not used)

derive_transport.make_nondivergent(w_in, v_in, cosc, cose, H, z, zm, y, dz, dy)

Ensure residual wind fields are non-divergence. This is lost through interpolation etc.

Parameters:

  • w_in (array) – Divergent w-wind.

  • v_in (array) – Divergent v-wind.

  • cosc (_type_) – Cosine of latitude at grid centres.

  • cose (_type_) – Cosine of latitude at grid edges.

  • H (float) – Scale height

  • z (array) – z-coordinate edge points in 2D model.

  • zm (array) – z-coordinate mid-points in 2D model.

  • y (array) – y-coordinate edge points in 2D model.

  • dz (array) – Grid cell heights.

  • dy (array) – Grid cell widths.

Returns:

Non-divergent w and v wind fields.

Return type:

arrays

derive_transport.add_convection(ds2d)

Add convective parameters straight from MERRA2 met data

Parameters:

ds2d (xarray dataset) – The monthly dataset of 2D transport to which the convection will be added

Returns:

A Data Array of the 2D monthly convective flux

Return type:

xarray dataarray

derive_transport.make_2D_yearly_files(start_year, end_year=None)

Take monthly 2D transport and make CF compliant 3D netcdf files

derive_transport.month