GNSS deep SNR retrievals of marine atmosphere boundary layer (MABL) specific humidity

This folder contains 5 prediction files ended with *_v2.h5. These files can be loaded using the provided code "prediction_data_loader.py". All variables are in their respective physical units. The *.tgz file contains the training and validation codes as well as sample training and validation datasets from METOP-B satellite. Please refer to the paper for details. All variables had been normalized in the training and validation datasets so no real physical meaning attached. Reference: Gong, J., Wu, D. L., Badalov, M., Ganeshan, M., and Zheng, M.: A machine-learning-based marine atmosphere boundary layer (MABL) moisture profile retrieval product from GNSS-RO deep refraction signals, Atmos. Meas. Tech., 18, 4025–4043, https://doi.org/10.5194/amt-18-4025-2025, 2025. POC: [e-mail] 10/17/2024 Update on 08/27/2025: The final paper has been published on AMT. Please see updated reference information above. --- THE END ---