pyaerocom is written and tested for Python >= 3.6 and for unix based systems. pyaerocom provides tools for processing and plotting of data related to the AeroCom project.

This includes reading and processing of gridded data (e.g. model data or level 3 satellite data, e.g. NetCDF files) and ungridded data (e.g. observational data from AERONET or EBAS networks, e.g. ASCII files) as well as tools for colocation and cross evaluation of different datasets.

Main features

  • Reading routines for many ground based observation databases, such as:

  • Reading routines for many space-based observations, such as:

  • Access to the AeroCom model database.

  • Data objects for analysis of gridded and ungridded (point-cloud) observations.

  • This includes interfaces for conversion of data to data types of related data analysis libraries such as pandas, numpy, xarray or iris.

  • Colocation tools for gridded and ungridded datasets.

  • Harmonisation of variable and metadata conventions.

  • Data visualisation tools and interfaces to common plotting libraries such as matplotlib or cartopy.

  • Tools for statistical analysis.

Usage examples


The AeroCom-project (http://aerocom.met.no/) is an open international initiative of scientists interested in the advancement of the understanding of the global aerosol and its impact on climate. A large number of observations (including MODIS, POLDER, MISR, AVHHR, SEAWIFS, TOMS, AATSR, AERONET and surface concentrations) and results from more than 14 global models have been assembled to document and compare state of the art modeling of the global aerosol. A common protocol has been established and models are asked to make use of the AeroCom emission inventories for the year 2000 and preindustrial times. Results are documented via interactive websites which give access to 2D fields and standard comparisons to observations. Regular workshops are held to discuss findings and future directions.

This repository contains the AeroCom python tools which are / will be used to produce the standard AeroCom analyses shown at the AeroCom phase 2 interface (http://aerocom.met.no/cgi-bin/AeroCom/aerocom/surfobs_annualrs.pl)

At this point the tools are co-operational together with the IDL based aerocom-tools that cannot be made public because they use 3rd party libraries with a non GPL compatible license.

Website and code documentation

The official website including code documentation is hosted here:



A list of all requirements is provided in file pyaerocom_env.yml.

Installing all requirements

NOTE: You can skip this section if you install the latest conda-forge release (more infos below under Installation).

We recommend using the Anaconda Python 3.7 distribution (or Miniconda, if you want to save disk space) and to use the conda package manager to install the requirements.

If you use Anaconda as a package manager, you can install all requirements (specified in previous section) into a new environment using the provided pyaerocom_env.yml file:

conda env create -n pya -f pyaerocom_env.yml

This will create a new conda environment called pya which can be activated using:

conda activate pya

Alternatively, you can include the requirements into an existing environment. First, activate the existing environment, and then install the dependencies using:

conda env update -f=pyaerocom_env.yml

Installation of pyaerocom

You have several options to install pyaerocom, the first one is the easiest, but may not refer to the most recent (non-released) version of pyaerocom. So please check first, which version you are interested in.

Option 1: Installation using conda install

NOTE: This will install the latest release of pyaerocom.

  • It hence, may not include the most recent developments.

  • Requirements are installed automatically.

If you use conda as a package manager, the easiest way to install pyaerocom (and all requirements, see previous section) is to use the build provided in the nordicesmhub conda channel:

conda install -c conda-forge pyaerocom

This will install the latest release of pyaerocom including all requirements. Alternatively, you may install from source as described in the following.

NOTE: installation support via conda as described above is quite recent, so please let us know if you run into problems with the installation (best way to do this is by raising an issue here).

Option 2: Install latest release via PyPi

NOTE: this will install the latest released version of pyaerocom, which is the same as distributed via conda-forge (see prev. point). However, installation via PyPi does not take care of any requirements but only installs pyaerocom:

pip install pyaerocom

Option 3: Installing from source

If you use the conda packages manager, please make sure to activate the environment you want to install pyaerocom into. For more information about conda environments, see here.

Please make sure to install all requirements (see above) before installing pyaerocom from source.

To install pyaerocom from source, please download and extract the latest release (or clone this repository) and install from the toplevel directory (that contains a file setup.py) using:

python setup.py install

Alternatively, if you plan to apply local changes to the pyaerocom source code, you may install in development mode:

python setup.py develop

You may also download and extract (or clone) the GitHub repo to install the very latest (not yet released) version of pyaerocom.

Access to users database

Please follow the instructions provided here, to retrieve access to the AeroCom users database:


NOTE: the users database does not contain any ground based observational data (such as EBAS, AERONET, etc.) but only the AeroCom model data available in the database as well as some gridded level 3 satellite datasets which may be used for model evaluation.

Getting started

After installing pyaerocom, open your python executable and try to import pyaerocom:

import pyaerocom as pya

To get started, please see introduction tutorial.

Tutorials (Jupyter notebooks)

A number of tutorial jupyter notebooks can be found in the pyaerocom-tutorials repo

NOTE: in order to execute these notebooks you need to be connected to the AeroCom database located on servers of the Norwegian Meteorological Institute (MetNo). Thus, if you are not happen to be an employee of the institute you will not be able to execute the tutorial notebooks on your own machine. We are working on providing an example dataset that can be downloaded and rewrite the notebooks based on these example data.

Remark for Windows users

pyaerocom is not tested on Windows systems and may only work in parts and thus some features may not work on Windows machines at the moment. In particular, features that rely on and are built upon access to the AeroCom database servers and automatic database path navigation. This includes the automised reading of gridded and ungridded data using either of the pre-defined path infrastuctures (e.g. check out paths.ini or paths_user_server.ini). However, you may still define file locations in your Python scripts yourself and use the more low-level features for reading the data. Windows support will be provided soon. Please let us know if you intend to use pyaerocom on a Windows machine so that we can consider adjusting our priorities, or also if you have any questions related to the usage.