Deep-learning Unsupervised Semantic Segmentation of GMI Satellite Brightness Temperature Images: Validation on precipitation and SST at Global and Regional Scale

Abstract Microwave brightness temperatures (TB) measured using satellites have played a significant role in observing various water cycle components. Until now, the measurement of TB has been used to retrieve many geophysical variables with Bayesian or, more recently, supervised Deep Learning approaches. These retrieval methods rely on the availability and the quality of the target variable and could be affected by the change in the TB and target variable relationship under climate change. This study aims to show that an unsupervised semantic image segmentation model on TB images offers meaningful segments that represent geophysical variables by interpreting and analyzing the resulting segmentation. The unsupervised segmentation methods explored include fully convolutional networks and pixel-wise approaches withK-means. The main challenge tackled here is evaluating the model’s performance without any ground truth labelling. The proposed evaluation requires external reference variables, including precipitation and sea surface temperatures (SST). The segmentation classes that represent these two geophysical variables are first compared to reference precipitation and SST data using segmentation metrics. The second step is to investigate the behaviour of these classes. Results show that the precipitation classes capture the pattern of monsoon precipitation, while the ocean classes can show the seasonality of SST. With an unsupervised approach, it is essential to verify the properties of the obtained segments extensively. The segmentation results in this study show that unsupervised learning is a key approach to extracting decades-long information in TB observations.