Open Science

For me open science means making results, code, and data usable and not just accessible – for the scientific community as well as society.

I try to follow best practices such as open access publishing and permissive licensing, coding style guides and documentation, metadata conventions and the FAIR data principles, as well as version control and transparency in data and code usage in general.

You are free to use any of the material I provide following the rules of the indicated licenses (mostly CC BY). In addition I’d be thankful if you let me know so that I can keep track (but you don’t have to).

If you publish scientific work based on material I provide please consider citing some of my papers and putting me in the acknowledgements if appropriate.


Running window bias: data and code accompanying Brunner et al 2024

In our 2024 paper we identify and discuss a bias merging from the use of too long seasonal windows when defining temperature extremes – the running window bias. The code and the most important datasets to recreate the figures from the paper and to play around with are freely available on GitHub.

Model Learning: Data and Code for identifying climate models

Model learning combines the terms climate model and machine learning providing a framework to disdinguish models and observations based on output maps. The backgrounds are described in Brunner and Sippel (2023), the data and code version used for the publication are available from Zotero, and the most recent version of the code can be found on GitHub.

Global blocking detection

During my PhD I implemented a blocking detection algorithm in Python and manly based on xarray. It enables the classification of atmospheric blocks based on global geopotential height fields and following different definitions from the literature. A detailed description can be found in section 3.2 (page 25f) of my PhD-thesis. The code is freely available under a MIT license on GitHub:

Climate model Weighting by Independence and Performance (ClimWIP) – Standalone

The Climate model Weighting by Independence and Performance (ClimWIP) package is based on earlier work by people from the Climate Physics group at ETH Zurich (mainly Reto Knutti, Jan Sedláček, and Ruth Lorenz). Ruth Lorenz and I have implemented the current version in Python and it is freely available under a GPLv3.0 on GitHub:

Climate model Weighting by Independence and Performance (ClimWIP) – ESMValTool

Supported by the eScience Center and the ESMValTool team (in particular Peter Kalverla) we have also added the ClimWIP method as a diagnostic to the ESMValTool, including several recipes to reproduce some of our published studies: ClimWIP recipes


CMIP6 next generation data archive

I have lead work to post-process CMIP6 data from ESGF into a quality-controlled, highly consistent database termed ETH Zurich CMIP6 next generation archive (CMIP6ng). Due to technical reasons and because we do not have dedicated funding for it the data is not directly available online. I do no longer work at ETH but they normally provide download options (ftp, rsync) for the data upon request: A documentation is available here.

Gridded radio occultation data

During my PhD I created a dataset of gridded radio occultation (RO) climatologies from 2006 to 2016. I used them to investigate atmospheric blocking and related impacts but they might also be useful of other applications. A detailed description of the data and the processing can be found in section 4.3 (page 35f) of my PhD-thesis. The data are freely available from the Climate Change Center Austria (CCCA):

  • geopotential height: link
  • temperature: link
  • specific humidity: link