Explaining neural networks for detection of tropical cyclones and atmospheric rivers in gridded atmospheric simulation data

Radke, Tim; Fuchs, Susanne; Wilms, Christian; Polkova, Iuliia; Rautenhaus, Marc

doi:https://doi.org/10.5194/gmd-18-1017-2025

Articles | Volume 18, issue 4

https://doi.org/10.5194/gmd-18-1017-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-18-1017-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 18, issue 4

Methods for assessment of models

|

24 Feb 2025

Methods for assessment of models |

| 24 Feb 2025

Explaining neural networks for detection of tropical cyclones and atmospheric rivers in gridded atmospheric simulation data

Tim Radke, Susanne Fuchs, Christian Wilms, Iuliia Polkova, and Marc Rautenhaus

Supplement

https://doi.org/10.5194/gmd-18-1017-2025-supplement

Data sets

ClimateNet Dataset as used in "Explaining neural networks for detection of tropical cyclones and atmospheric rivers in gridded atmospheric simulation data" Tim Radke https://doi.org/10.5281/zenodo.14046402

Model code and software

Code for the paper: "Explaining neural networks for detection of tropical cyclones and atmospheric rivers in gridded atmospheric simulation data" Tim Radke et al. https://doi.org/10.5281/zenodo.10892412

Short summary

In our study, we built upon previous work to investigate the patterns artificial intelligence (AI) learns to detect atmospheric features like tropical cyclones (TCs) and atmospheric rivers (ARs). As primary objective, we adopt a method to explain the AI used and investigate the plausibility of learned patterns. We find that plausible patterns are learned for both TCs and ARs. Hence, the chosen method is very useful for gaining confidence in the AI-based detection of atmospheric features.