Login Sign up

Improving reproducibility of hydro-meteorological research and publications with Metview

Fatima Pillosu, Iain Russell

Audience level:

Brief Summary

Metview (a software for accessing, processing and visualising specialized meteorological data) has been recently integrated with Jupyter notebooks and Binder. The poster shows how Metview is used within a PhD work @ReadingUni & @ECWMF to create fully reproducible scientific workflows and publications.


Nowadays, the goal of raising scientific research and publication standards by creating fully reproducible workflows and publications have been gaining momentum in world-leading research institutions and scientific journals like Nature and Science. In recent years, some attempts to develop research and write publications that comply with high reproducibility standards has been made also within the hydro-meteorological community. However, hydro-meteorological scientists are not generally computational experts. Still, they find themselves using an increasingly diverse set of languages, software tools, packages, datasets, etc. to develop their research and write their publications. Furthermore, the majority of researchers in the field have acquired the knowledge required to develop and use scientific software primarily from peers or through self-study, rather than from formal education and training. Therefore, they might not know how to (1) adequately specify the software packages and libraries used to develop their analysis and expose their results, nor (2) how to make data and code available in the most efficient way. Therefore, it might be difficult to share their knowledge and make their publications available to peers who wish to validate techniques or results. Metview is a workstation application aimed at both the operational and research meteorologist. Its capabilities include meteorological data access, processing and visualisation. It features a rich library of Python functions and classes, and a powerful user interface for interactive work. This poster will show how Metview has been recently adapted to be used within tools like Zenodo, Jupyter notebooks and Binder to help hydro-meteorologists to share their science complying with the highest standards in workflows and publications reproducibility. The poster will also show a real-world application within a PhD, hoping that this experience will encourage more students to start their careers following this path.