Abstract

North Africa is the world’s largest source region of mineral dust. Mineral dust aerosol itself plays an important role in the climate system, as it is, for example, directly and indirectly influencing radiative transfer and providing nutrients for marine and terrestrial ecosystems. In addition, airborne mineral dust has adverse effects on air quality and public health. Satellite observations can provide large spatial coverage of dust plumes, which facilitates the study of dust sources, transport pathways, and sinks. Such large spatial coverage can be combined with a high temporal resolution by instruments onboard geostationary satellite. An example of such an instrument is the Spinning Enhanced Visible and InfraRed Imager (SEVIRI) onboard the geostationary Meteosat Second Generation satellites (MSG). The full spatial extent of dust plumes in SEVIRI observations is frequently obscured by clouds. To overcome this limitation, we propose the use of machine-learning-based image in-painting techniques. Machine-learning-based image in-painting techniques can restore damaged images of structures like buildings, cars, landscapes, insects or human faces by learning the typical patterns of these structures. Image in-painting algorithms have in recent years been successfully adapted to reconstruct missing geophysical data. In this study, we use an off-the-shelf implementation of an image in-painting algorithm and developed a method for applying it to satellite-observed dust plumes. The algorithm is trained on reanalysis fields of the dust aerosol optical thickness combined with temporally corresponding cloud masks obtained from MSG-SEVIRI. In a next step we use this trained algorithm to restore the full spatial extent of dust plumes on grey-scaled images of North African dust plumes during 2021 and 2022, derived from the SEVIRI Dust RGB product. We test the reconstructed dust plumes against independent data, derived from dust forecasts provided by the WMO Barcelona Dust Regional Center. Our reconstructions spatially and temporally agree well with output from the forecast model ensemble. The proposed method is adaptable to other satellite products in the future, including products from the Meteosat Third Generation Flexible Combined Imager (MTG-FCI). Reference Kanngießer and Fiedler, 2024, “Seeing” beneath the clouds - machine-learning-based reconstruction of North African dust plumes, AGU Advances, In Press.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.