Detection and analysis of anomalies in the brightness temperature difference field using MSG rapid scan data

The paper addresses anomalies in brightness temperature difference (BTD) between WV and IR window bands using MSG rapid scan data. One of our aims is to eliminate subjectivity in their detection. The important finding is that it is not possible to study the anomalous behavior when taking into account all pixels related to a storm cloud top. It is necessary to identify and study anomalous regions separately, otherwise their behavior is masked by the majority of the other pixels. We propose a detection algorithm based on the thresholding method with the threshold depending on the dispersion of the BTD data. Outputs of the algorithm and first analyses for one chosen case are shown. According to the first results, the typical relationship between BT and BTD (most of the pixels with growing BTD exhibit a decrease in BT) is retained during the time evolution of the storm also for anomalous pixels, even though their BTDs reach higher values than for non-anomalous pixels at a similar BT level.

► We propose an algorithm for objective detection of the BTD anomalies.
► To study anomalous behaviour of BTD it is necessary to select anomalous pixels.
► Most of the pixels with growing BTD exhibit a BT decrease also for anomalous pixels.
► BT and BTD for anomalous pixels reach extremes when the number of pixels culminates.