نبذة مختصرة : Australian dust storms have principally been studied using ground observations that are limited in being spatially sparse, and low Earth orbit satellites in turn limited by their relatively poor temporal resolution. In this study, the high temporal resolution of geostationary satellite monitoring was harnessed to assess the ability of Himawari-8 Advanced Himawari Imager (AHI) at detecting and mapping five case study Australian dust storm events between 2019 and 2020. AHI thermal bands (9μm,10μm and 11μm) were used to create Brightness Temperature Difference (BTD) indices which were then applied to a Threshold-based Dust Detection Algorithm (TBDDA). Event-specific index thresholds were tuned for each event at the time of the highest dust concentration as recorded by a ground observation network. Selection of event-specific thresholds showed considerable variation and subsequent evaluation of dust storm detection revealed a low average Probability of Detection (POD) of 0.01 and a False Alarm Ratio (FAR) of 0.3 for the five events. The variability within the thresholds for each index across the five dust storms emphasises the lack of a singly suitable threshold set for event detection. It highlights inefficiency in the TBDDA when observing individual Australian dust events. A method that does not rely on event-specific thresholds is seen as the next step in advancing dust storm detection in Australia.
Processing Request
No Comments.