How to install nctoolkit

You will need a Linux or Mac operating system for nctoolkit to work. It will not work on Windows due to system requirements.

The best and easiest way to install nctoolkit is to use conda. This will install all system dependencies, and nctoolkit will just work out of the box. This can be done as follows:

$ conda install -c conda-forge nctoolkit
Conda can often install a very old version of nctoolkit. So you might want to install the latest version. You can find the latest version available on conda

here. If you wanted to install version version 1.1.6, you would do the following:

$ conda install -c conda-forge nctoolkit=1.1.6

Mamba is a smoother way to manage conda environments. If you don’t use it, you should try. Install it from here.

Once mambaforge is installed you can install nctoolkit as follows:

$ mamba install -c conda-forge nctoolkit

This will be much faster to install than using conda, because mamba resolves environments much faster.

Note that recent releases are not available on macOS on conda. This issue is being investigated at the minute, and will hopefully be resolved shortly. In the meantime, if you are using macOS, it is best to install using pip.

If you do not use conda, you can install nctoolkit using pip. The package is available from the Python Packaging Index. To install nctoolkit using pip:

$ pip install nctoolkit

This will install the core dependencies of nctoolkit. If you want slightly better plots, you can install all dependencies:

$ pip install nctoolkit[complete]

Once you have installed nctoolkit using pip, you will need to install the system dependencies listed below.

To install the development version from GitHub:

$ pip install git+

Known issues

Recent updates to dependencies may cause issues with certain versions of nctoolkit. Please ensure you are using up to date versions of geoviews if you want to use coastlines in plots, and jupyter_bokeh if you want to use plots in jupyter notebooks.

To do this, run the following commands:

$ conda install -c conda-forge geoviews=1.10.1

$ conda install -c conda-forge jupyter_bokeh=3.0.7

This issue should not impact fresh installs of nctoolkit, which will have dependency pinning.

How to get better looking plots

If you have installed the non-complete version of nctoolkit from pypi, you might also want to install cartopy to get better looking plots. This has some additional dependencies, so you may need to follow their guide here to ensure cartopy is installed fully.

If you installed nctoolkit using conda, you can install cartopy as follows:

$ mamba install -c conda-forge cartopy

$ conda install -c conda-forge cartopy

Cartopy is not installed by default because installation can be slow or difficult on pypi or conda. However, no such issues exist on mamba. So, use mamba.

Python dependencies

System dependencies

There are two main system dependencies: Climate Data Operators, and NCO. The easiest way to install them is using conda:

$ conda install -c conda-forge cdo

$ conda install -c conda-forge nco

or mamba:

$ mamba install -c conda-forge cdo

$ mamba install -c conda-forge nco

CDO is necessary for the package to work. NCO is an optional dependency and does not have to be installed.

If you are working on an Ubuntu system, you should be able to install CDO as follows:

$ sudo apt install cdo

If you want to install CDO from source, you can use one of the bash scripts available here.