pyaerocom is an open-source software for the AeroCom project, written in Python 3. It is being developed at the Norwegian Meteorological Institute since 2018, and it is continuously improved and extended in the supported functionality.
pyaerocom provides both high-level and low-level processing routines allowing for highly customizable model assessments. The high-level tools are used for semi-automated assessments of model data which are displayed in the Phase III evaluation interface.
The methods and tools provided by the lower-level pyaerocom API have been successfully used by many Bachelor, Master and PhD students in several University courses focusing on assessments of models and inter-comparison with observations, such as the 2018 NEGI eScience course in Andøya, northern Norway, the 2019 NEGI eScience course in Abisko, northern Sweden, and the 2020 FORCeS eSciences course, which was held fully online. An introduction into the API is provided in the pyaerocom tutorials, which can also be run interactively using Binder.
pyaerocom supports reading routines for several commonly used ground-based observational databases and formats, such as AERONET, EBAS, EARLINET. In addition, it supports reading routines for level 3 gridded satellite data as well lower level point-cloud satellite data, for instance, for Sentinel-5P or Aeolus.
These observational records can be readily used for the assessment of model simulations, through sophisticated co-location routines implemented in pyaerocom.
pyaerocom makes large use of other open-source software projects, which are highly acknowledged. These include, for instance, basic python libraries such as numpy, pandas, matplotlib or cartopy, but also libraries for model data analysis such as iris or xarray.
Github repository: https://github.com/metno/pyaerocom
Tutorials Github repository: https://github.com/metno/pyaerocom-tutorials
Interactive version of tutorials: https://mybinder.org/v2/gh/metno/pyaerocom-tutorials/master