Articles | Volume 19, issue 8
https://doi.org/10.5194/gmd-19-3213-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/gmd-19-3213-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
MeteoSaver v1.0: a machine-learning based software for the transcription of historical weather data
Vrije Universiteit Brussel, Department of Water and Climate, 1050 Brussels, Belgium
Bas Vercruysse
Ghent University, Department of History, Ghent Centre for Digital Humanities, 9000 Ghent, Belgium
Krishna Kumar Thirukokaranam Chandrasekar
Ghent University, Department of History, Ghent Centre for Digital Humanities, 9000 Ghent, Belgium
Royal Museums of Art and History, 1000 Brussels, Belgium
Christophe Verbruggen
Ghent University, Department of History, Ghent Centre for Digital Humanities, 9000 Ghent, Belgium
Julie M. Birkholz
Ghent University, Department of History, Ghent Centre for Digital Humanities, 9000 Ghent, Belgium
Digital Research Lab, KBR – Royal Library of Belgium, 1000 Brussels, Belgium
Koen Hufkens
BlueGreen Labs (bv), 9120 Melsele, Belgium
Hans Verbeeck
Ghent University, Department of Environment, 9000 Ghent, Belgium
Pascal Boeckx
Ghent University, Isotope Bioscience Laboratory – ISOFYS, 9000 Ghent, Belgium
Seppe Lampe
Vrije Universiteit Brussel, Department of Water and Climate, 1050 Brussels, Belgium
Ed Hawkins
University of Reading, National Centre for Atmospheric Science, Department of Meteorology, RG6 6ET Reading, UK
Peter Thorne
Maynooth University, ICARUS Climate Research Centre, Maynooth, Ireland
Dominique Kankonde Ntumba
Institut National pour l'Etude et la Recherche Agronomiques, Direction Générale, Kinshasa, Democratic Republic of the Congo
Olivier Kapalay Moulasa
Institut National pour l'Etude et la Recherche Agronomiques, Direction Générale, Kinshasa, Democratic Republic of the Congo
Wim Thiery
Vrije Universiteit Brussel, Department of Water and Climate, 1050 Brussels, Belgium
Model code and software
MeteoSaver v1.0 (v1.0-final) Derrick Muheki et al. https://doi.org/10.5281/zenodo.19123862
Short summary
Archives worldwide host vast records of observed weather data crucial for understanding climate variability. However, most of these records are still in paper form, limiting their use. To address this, we developed MeteoSaver, an open-source tool, to transcribe these records to machine-readable format. Applied to ten handwritten temperature sheets, it achieved a median accuracy of 74 %. This tool offers a promising solution to preserve records from archives and unlock historical weather insights.
Archives worldwide host vast records of observed weather data crucial for understanding climate...