Articles | Volume 19, issue 11
https://doi.org/10.5194/amt-19-3865-2026
https://doi.org/10.5194/amt-19-3865-2026
Research article
 | 
12 Jun 2026
Research article |  | 12 Jun 2026

Estimation of Doppler velocity from incoherent scatter spectra using context-aware transformers

Yanlin Li and Qihou Zhou

Related authors

Variance estimations in the presence of intermittent interference and their applications to incoherent scatter radar signal processing
Qihou Zhou, Yanlin Li, and Yun Gong
Atmos. Meas. Tech., 17, 4197–4209, https://doi.org/10.5194/amt-17-4197-2024,https://doi.org/10.5194/amt-17-4197-2024, 2024
Short summary

Cited articles

Aponte, N., Sulzer, M. P., Nicolls, M. J., Nikoukar, R., and Gonzalez, S. A.: Molecular ion composition measurements in the F1 region at Arecibo, J. Geophys. Res.-Space, 112.A6, https://doi.org/10.1029/2006JA012028, 2007. 
Ba, J. L., Kiros, J. R., and Hinton, G. E.: Layer normalization, arXiv [preprint], https://doi.org/10.48550/arXiv.1607.06450, 2016. 
Chau, J., Fejer, B., and Goncharenko, L.: Quiet variability of equatorial E × B drifts during a sudden stratospheric warming event, Geophys. Res. Lett., 36, https://doi.org/10.1029/2008gl036785, 2009. 
Chau, J. L. and Kudeki, E.: First E- and D-region incoherent scatter spectra observed over Jicamarca, Ann. Geophys., 24, 1295–1303, https://doi.org/10.5194/angeo-24-1295-2006, 2006. 
Devlin, J., Chang, M. W., Lee, K., and Toutanova, K.: BERT: Pre-training of deep bidirectional transformers for language understanding, Proc. 2019 Conf. of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), 4171–4186, https://doi.org/10.18653/v1/N19-1423, 2019. 
Download
Short summary
We introduce a transformer-based AI model for estimating Doppler velocity from incoherent scatter radar (ISR) spectra. Inspired by Vision Transformers, the model uses a standard transformer encoder adapted for radar data. Trained solely on simulated spectra, it performs well on real data from the Arecibo radar and significantly outperforms the traditional least-squares fitting (LSF) method. This approach is potentially applicable wherever spectral data can be parameterized.
Share