Developing a Neural Network for the Identification of EMIC Wave Events

A supervised convolutional neural network (CNN) was developed to automatically identify electromagnetic ion cyclotron (EMIC) wave events from spectrograms. These events have normally been identified manually, which can be a time consuming process. Statistical analyses of larger datasets would be facilitated if this extraction process could be simplified. The neural network model was trained on spectrogram images from the Halley magnetometer station that had been manually identified as either containing or not containing an EMIC wave event anywhere in the spectrogram. This model was then tested on an unseen set of spectrograms, resulting in a reasonable classification accuracy (true positive rate > 95%). Additionally, both the frequency and time frame of the events in each spectrogram were extracted from the model. This method has the capability of reducing the time and effort required to identify important spectrogram features by hand. Such an automated method could also be applied to other space weather data stored in spectrograms.