EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Addressing the Labelled Observation Data Gap for Biological and Meteorological Radar Echo Classification: A Transfer Learning Framework Using NEXRAD Data

Fidele Maniraguha, Anthony Vodacek, Vedaste Iyakaremye, Aimable Gahigi, Rija Faniriantsoa, Germain Nkima

IET Radar, Sonar & Navigation · 2026-01 · 20 (1)

DOI: 10.1049/rsn2.70261 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IET Radar, Sonar & Navigation, online erschienen am 30.10.2026. Verfasst von Fidele Maniraguha und 5 weiteren Autor:innen. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Discriminating biological from meteorological radar echoes is essential for precipitation estimation, data quality control and ecological monitoring. In tropical regions, such as Rwanda, no labelled biological reference data exist to calibrate standard algorithms. We adapt three techniques to Rwanda's C‐band dual‐polarisation radar and quantify their transferability from S‐band NEXRAD: the physics‐based depolarisation ratio (DR) method, MistNet, a U‐Net convolutional network and RadxPID, a fuzzy‐logic hydrometeor classifier. Four volumes are evaluated: S‐band mixed echo and clear‐air migration, C‐band fall armyworm season and light rain. Skill is measured using fraction of echoes identified (FEI) and Heidke skill score (HSS); cross‐band transfer by distribution overlap (OVL), Hellinger distance and Cohen's kappa. On S‐band, DR reaches FEI = 75.3% (meteorological) and 99.4% (biological) with HSS = 0.75, exceeding MistNet (0.66) and RadxPID (0.52). Across bands, ρHV transfers best (OVL = 0.76) whereas ZH shifts most (OVL = 0.45), explaining why DR degrades gracefully at C‐band whereas MistNet's biological classification collapses below 2% of gates. Biological classification follows the boundary layer with range. Unanimous inter‐method agreement is 50%–61% on S‐band and 55% on C‐band. Physics‐based methods retain cross‐band coverage; data‐driven methods need retraining. DR is the most deployable option for data‐limited tropical environments.

Zitieren

APA

Maniraguha, F., Vodacek, A., Iyakaremye, V., Gahigi, A., Faniriantsoa, R., & Nkima, G. (2026). Addressing the Labelled Observation Data Gap for Biological and Meteorological Radar Echo Classification: A Transfer Learning Framework Using NEXRAD Data. IET Radar, Sonar & Navigation, 20(1). https://doi.org/10.1049/rsn2.70261

Vancouver

Maniraguha F, Vodacek A, Iyakaremye V, Gahigi A, Faniriantsoa R, Nkima G. Addressing the Labelled Observation Data Gap for Biological and Meteorological Radar Echo Classification: A Transfer Learning Framework Using NEXRAD Data. IET Radar, Sonar & Navigation. 2026;20(1). doi:10.1049/rsn2.70261

BibTeX
@article{Maniraguha2026,
  author = {Maniraguha, Fidele and Vodacek, Anthony and Iyakaremye, Vedaste and Gahigi, Aimable and Faniriantsoa, Rija and Nkima, Germain},
  title = {Addressing the Labelled Observation Data Gap for Biological and Meteorological Radar Echo Classification: A Transfer Learning Framework Using NEXRAD Data},
  journal = {IET Radar, Sonar & Navigation},
  year = {2026},
  volume = {20},
  number = {1},
  doi = {10.1049/rsn2.70261}
}
Kostenlos starten →