Sci Rep. 2026 Jun 3;16(1):17165. doi: 10.1038/s41598-026-55422-y.
ABSTRACT
On 15 March 2022 an intense, but partially cloud-obscured, Saharan dust plume was transported towards Europe by an atmospheric river. Ten-year statistics of dust plumes co-occurring with cloud cover over Europe highlights that regionally up to 100% of dust plumes are obscured by clouds, which poses challenges for reconstructing dust plumes from satellite images. The European dust case on 15 March 2022 is used to investigate whether data-driven machine-learning techniques for restoring the spatial extent of dust plumes in SEVIRI satellite images can be alleviated by exploiting the rich ground-based data sets in Europe. Satellite images were paired with different combinations of ground-based observation data, stemming from ground-based remote sensing, weather reports and measurements of particulate matter, using a k-nearest neighbours approach. Combining ceilometer and photometer data with satellite images added the most value for restoring the dust plume extent and are recommended for future reconstructions of cloud-obscured dust plumes.
PMID:42236821 | DOI:10.1038/s41598-026-55422-y
