gsw-xarray: Wrapper for gsw that adds CF attributes

gsw-xarray is a wrapper for gsw python that will add CF attributes to xarray.DataArray outputs. It is meant to be a drop in wrapper for the upstream GSW-Python library and will only add these attributes if one argument to a function is an xarray.DataArray.

You can find the documentation on gsw-xarray.readthedocs.io.

Usage

Simple usage

import gsw_xarray as gsw

# Create a xarray.Dataset
import numpy as np
import xarray as xr
ds = xr.Dataset()
id = np.arange(3)
ds['id'] = xr.DataArray(id, coords={'id':id})
ds['CT'] = ds['id'] * 10
ds['CT'].attrs = {'standard_name':'sea_water_conservative_temperature'}
ds['SA'] = ds['id'] * 0.1 + 34
ds['SA'].attrs = {'standard_name':'sea_water_absolute_salinity'}

# Apply gsw functions
sigma0 = gsw.sigma0(SA=ds['SA'], CT=ds['CT'])
print(sigma0.attrs)

Outputs

{'standard_name': 'sea_water_sigma_t', 'units': 'kg/m^3'}

Don't worry about usage with non xarray array objects, just use in all places you would the upstream library:

sigma0 = gsw.sigma0(id * 10, id * 0.1 + 34)
print(type(sigma0), sigma0)

Outputs

<class 'numpy.ndarray'> [-5.08964499  2.1101098   9.28348219]

Using Pint

We support (and convert the unit if necessary) the usage of pint.Quantities and the usage of xarray wrapped Quantities. Support for pint requires the installation of two optional dependencies: pint and pint-xarray. If all the inputs to a gsw function are Quantities, the returned object will also be a Quantity belonging to the same UnitRegistry.

Warning

Quantities must all belong to the same pint.UnitRegistry, a ValueError will be thrown if there are mixed registries.

Warning

If one input is a Quantity, all inputs must be Quantities (and/or xarray wrapped Quantities), except for the axis and interp_method arguments. For mixed usage of Quantities and non Quantities, a ValueError will be thrown.

import pint_xarray
import gsw_xarray as gsw

# Create a xarray.Dataset
import numpy as np
import xarray as xr
ds = xr.Dataset()
id = np.arange(3)
ds['id'] = xr.DataArray(id, coords={'id':id})
ds['CT'] = ds['id'] * 10
# make sure there are unit attrs this time
ds['CT'].attrs = {'standard_name':'sea_water_conservative_temperature', 'units': 'degC'}
ds['SA'] = ds['id'] * 0.1 + 34
ds['SA'].attrs = {'standard_name':'sea_water_absolute_salinity', 'units': 'g/kg'}

# use the pint accessor to quantify things
ds = ds.pint.quantify()

# Apply gsw functions
sigma0 = gsw.sigma0(SA=ds['SA'], CT=ds['CT'])
# outputs are now quantities!
print(sigma0)

Outputs

<xarray.DataArray 'sigma0' (id: 3)>
<Quantity([27.17191038 26.12820162 24.03930887], 'kilogram / meter ** 3')>
Coordinates:
  * id       (id) int64 0 1 2
Attributes:
    standard_name:  sea_water_sigma_t

The usage of xarray wrapped Quantities is not required, you can use pint directly (though the pint-xarray dep still needs to be installed).

import gsw_xarray as gsw
import pint
ureg = pint.UnitRegistry()
SA = ureg.Quantity(35, ureg("g/kg"))
CT = ureg.Quantity(10, ureg.degC)
sigma0 = gsw.sigma0(SA=SA, CT=CT)
print(sigma0)

Outputs

26.824644457868317 kilogram / meter ** 3

As gsw-xarray converts arguments to the proper unit when Quantities are used, we can e.g. use the temperature in Kelvin:

CT = ureg.Quantity(10, ureg.degC).to('kelvin')
sigma0 = gsw.sigma0(SA=SA, CT=CT)
print(sigma0)

Outputs

26.824644457868317 kilogram / meter ** 3

Note

If you do not wish to use the unit conversion ability, you need to pass dequantified Quantities (e.g. da.pint.dequantify() for pint-xarray or arg.magnitude for pint.Quantity).

Warning

On the opposite, gsw-xarray will not check the units if non Quantity arguments are used. If you wish to use unit conversion, please pass quantified arguments (if your xarray.Dataset / xarray.DataArray has the 'units' attribute, you can use da.pint.quantify())

Note

We recommend that you use the cf-xarray registry for units, as it implements geophysical units as degree_north, degrees_north, etc. gsw-xarray internally uses degree_north and degree_east for latitude and longitude unit names. If they are not found in the unit registry, they will be replaced by degree.

The function gsw.SP_from_SK uses part per thousand for SK. 'ppt' is already used for picopint, so the expected unit is replaced by '1'.

Xarray accessor

gsw-xarray provides a new accessor for xarray, that allows to call the gsw functions directly on a dataset:

ds.gsw.sigma0(CT="CT", SA="SA")
# or
ds.gsw.sigma0(CT=ds.CT, SA=ds.SA)
# or even, if CT and SA have the proper standard names
ds.gsw.sigma0()

Any type of mixte usage with dataArrays, numbers, strings, or autoparse with standard names is possible.

If all arguments are present in the dataset with the proper standard name, it is possible to use the accessor with brackets, as if it was a dictionary

ds.gsw["sigma0"]
# Or if you want to get a list of multiple variables
ds.gsw[["sigma0", "alpha"]]

If the dataset contains multiple variables with same standard name (e.g. practical salinity from bottle or CTD), you can set an option to tell gsw_xarray which variable to get:

# Globally
gsw_xarray.set_cf_name_preference(standard_name="variable_in_dataset")
# Or in a context, e.g.
with gsw_xarray.set_non_cf_name(sea_water_pressure="pres_adjusted"):
    # write code here
    pass

If you wish to use the accessor with automatic detection of arguments, but for a function whose arguments do not have a standard name, it is possible. You need to set an option in gsw-xarray, either in a context or globally

# Globally
gsw_xarray.set_non_cf_name(argument="argument_name_in_dataset")
# Or in a context, e.g.
with gsw_xarray.set_non_cf_name(Rt="Rt_in_ds"):
    ds.gsw.SP_salinometer(t=0)

In this 2nd case, the function gsw.SP_salinometer take the argument Rt which has no standard name.

When using user set options, the order of priority to automatically get variables is: 1) variables set by gsw_xarray.set_cf_name_preference, 2) variables with standard name (internal mapping), and 3) variables set by gsw_xarray.set_non_cf_name.

Installation

Pip

pip install gsw-xarray

Conda

Inside a conda environment: conda install -c conda-forge gsw-xarray.

Citation

If you use gsw-xarray, please cite the reference paper for the upstream gsw library: McDougall, T.J. and P.M. Barker, 2011: Getting started with TEOS-10 and the Gibbs Seawater (GSW) Oceanographic Toolbox, 28pp., SCOR/IAPSO WG127, ISBN 978-0-646-55621-5

@book{mcdougall2011getting,
  author = {McDougall, T. J. and Barker, P. M.},
  title = {Getting Started with TEOS-10 and the Gibbs Seawater (GSW) Oceanographic Toolbox},
  year = {2011},
  pages = {28},
  publisher = {SCOR/IAPSO WG127},
  isbn = {978-0-646-55621-5}
}

You can also cite gsw-xarray by using the zenodo DOI .

Contributor guide

All contributions, bug reports, bug fixes, documentation improvements, enhancements, and ideas are welcome. If you notice a bug or are missing a feature, fell free to open an issue in the GitHub issues page.

In order to contribute to gsw-xarray, please fork the repository and submit a pull request. A good step by step tutorial for starting with git can be found in the xarray contributor guide. A main difference is that we do not use conda as python environment, but poetry.

Set up the environment

You will first need to install poetry. Then go to your local clone of gsw-xarray and launch installation:

cd /path/to/your/gsw-xarray
poetry install --with dev

You can then activate the environment by launching a shell within the virtual environment:

poetry shell

You can check that the tests pass locally:

pytest gsw_xarray/tests

You can install pre-commit to run the linting automatically at each commit.

Release (for maintainers only)

TODO...