Gravity-wave-detecting CNN and JAGUAR-as-AIRS dataset

Processed dataset shown in the figures of Okui et al. (2025; submitted to Geophysical Research Letters on 28 February 2025). The Japanese Atmospheric General circulation model for Upper Atmosphere Research (JAGUAR) is a T639 global spectral model with high vertical resolution (300 m in the middle atmosphere) that simulates the entire neutral atmosphere (Watanabe & Miyahara, 2009; Okui et al., 2021). This dataset provides JAGUAR simulation outputs resampled as temperatures retrieved from Atmospheric Infrared Sounder (AIRS) measurements (Hoffmann & Alexander, 2009). A wave-detecting convolutional neural network (CNN) applicable for nadir-viewing satellite observations and numerical simulation outputs. CNN_run_AIRS.py is a sample code to run the CNN. For details, please refer to the paper.