i7aof.biascorr¶

Bias-correction utilities and workflows. See i7aof.biascorr for developer details.

class i7aof.biascorr.timeslice.Timeslice(config, thetao, so, basinmask, basinNumber, yidx=None)¶

Bases: object

A class for a time slice containing T and S

config¶

Configuration options.

Type:: mpas_tools.config.MpasConfigParser

thetao¶

name of file containing thetao

Type:: str

so¶

name of file containing so

Type:: str

Public¶

------

V¶

gridded volume in 3D space

Type:: float([nz, ny, nx])

T¶

gridded temperature in 3D space

Type:: float([nz, ny, nx])

S¶

gridded salinity in 3D space

Type:: float([nz, ny, nx])

Vb¶

binned volume in S, T space per basin

Type:: float([Nbasins, Nbins, Nbins])

Sb¶

Evenly spaced salinity bins per basin

Type:: float([Nbasins, Nbins + 1])

Tb¶

Evenly spaced temperature bins per basin

Type:: float([Nbasins, Nbins + 1])

deltaSf¶

Filled delta S between this time slice and the model reference, in S, T space per basin

Type:: float([Nbasins, Nbins, Nbins))

deltaTf¶

Filled delta T between this time slice and the model reference, in S, T space per basin

Type:: float([Nbasins, Nbins, Nbins))

S_corrected¶

Bias-corrected salinity in 3D space

Type:: float([nz, ny, nx])

T_corrected¶

Bias-corrected temperature in 3D space

Type:: float([nz, ny, nx])

Sperc¶

Requested percentile of salinity, usually 99th

Type:: float([Nbasins])

Tperc¶

Requested percentile of temperature, usually 99th

Type:: float([Nbasins])

Sscaling¶

Scaling factor of salinity, effectively stretching or compressing the model salinity distribution to best fit the reference distribution whilst fixing Sperc. Only values between 0.5 and 1.5 are allowed.

Type:: float([Nbasins])

Tscaling¶

Scaling factor of temperature, effectively stretching or compressing the model temperature distribution to best fit the reference distribution whilst fixing the minimum temperature value (usually freezing point).

Type:: float([Nbasins])

apply_anomaly(base)¶

Determine model anomaly with respect to reference period and apply anomaly to the base T and S to get the corrected values

Parameters:: base (Timeslice) – base timeslice on which to add T- and S-anomalies. This should usually be the timeslice of the same model over the reference period (same as modref), but with the extrapolated fields, so that T_corrected and S_corrected are also extrapolated over the full domain

compute_delta(modref)¶

Compute difference between timeslice and modref, binned on the modelref bins

Parameters:: modref (Timeslice) – Timeslice of the same model as this Timeslice over the reference period. This should usually be the non-extrapolated fields

fill_delta(deltavar)¶

Fill deltavar (either deltaT or deltaS) in full normalised T,S space

Parameters:: deltavar (array([Nbins, Nbins])) – deltaT or deltaS in S,T space

get_T_and_S()¶: Extract T and S data

get_all_data()¶: Extract all data from time slice

get_bins()¶: Get 2D histogram of volume Vb as a function of binned salinity (Sb) and temperature (Tb)

get_volume()¶: Extract volume per grid cell

class i7aof.biascorr.projection.Projection(config, logger)¶

Bases: object

A main class for projections

config¶

Configuration options.

Type:: mpas_tools.config.MpasConfigParser

logger¶

Logger for the class.

Type:: logging.Logger

Public¶

------

ref¶

Reference timeslice to which a model will be bias-corrected. This should usually be an observational dataset

Type:: Timeslice

modref¶

Model reference timeslice. This should be a model Timeslice averaged over the reference period. The input fields (T,S) should not be extrapolated, to prevent biases in the volume of certain water masses.

Type:: Timeslice

base¶

Base timeslice to which anomalies are added. This should be a model Timeslice over the reference period, similar to modref. Usually, the base Timeslice should contain extrapolated T, S fields though, to ensure the output bias-corrected fields are also extrapolated.

Type:: Timeslice

basinNumber¶

Gridded product denoting the individual basins by integer values Usually, this should be the 16 IMBIE basins, though a modification of these is required to ensure each contains a finite volume over the continental shelf. This same modification is required for the basic extrapolation of T and S.

Type:: int([ny, nx])

basinmask¶

Mask containing 0s and 1s, denoting which horizontal grid cell is part of which basin. 1: True, 0: False

Type:: int([Nbasins, ny, nx])

compute_S_bias(**kwargs)¶

compute_T_bias(**kwargs)¶

compute_bias()¶: Compute the bias of the model T and S with respect to the reference. This will create corrected T and S bins (Tc and Sc)

create_basin_masks(**kwargs)¶

get_model_info()¶: Extract model info from config file

plot_TS_diagrams(**kwargs)¶

read_model()¶: Read whole model period

read_model_timeslice(thetao_file, so_file, yidx)¶: Read a timeslice from the future period

read_reference(**kwargs)¶

i7aof.biascorr.projection.status(process_name)¶