The Convection Initiation Nowcasting Deep Learning Dataset v1.0

The GOES-16 Convective Initiation dataset version 1.0 contains channel 8-14 infrared brightness temperatures at lead times from 60 minutes to 10 minutes every 10 minutes over the Great Plains area. These data are used to train and test a physically explainable deep learning neural network to predict convection initiation. The data is stored in local patches surrounding each convection initiation (CI), near miss (NM), and randomly selected (RM) event. The 'train' directory contains the training dataset from June and July in 2020 while the 'test' directory contains the testing dataset from June in 2021. This dataset is downloadable as netCDF files, and is documented in a companion paper: Fan, D., S. J. Greybush, D. J. Gagne, and E. E. Clothiaux, 2023: Physically Explainable Deep Learning for Convective Initiation Nowcasting Using GOES-16 Satellite Observations.