Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data

Accurate cloud cover assessment is crucial in several fields, such as weather forecasting, climate science, agriculture, or energy system planning, precipitation pattern forecasting and aiding in early detection of extreme weather events. Despite the crucial data that weather stations provide about...

Full description

Saved in:
Bibliographic Details
Main Authors: Markus Gotz, Erik Wessel, Ugur Cayoglu, Babak Jahani, Charlotte Debus, Jan Cermak, Achim Streit
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10949748/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849713198526103552
author Markus Gotz
Erik Wessel
Ugur Cayoglu
Babak Jahani
Charlotte Debus
Jan Cermak
Achim Streit
author_facet Markus Gotz
Erik Wessel
Ugur Cayoglu
Babak Jahani
Charlotte Debus
Jan Cermak
Achim Streit
author_sort Markus Gotz
collection DOAJ
description Accurate cloud cover assessment is crucial in several fields, such as weather forecasting, climate science, agriculture, or energy system planning, precipitation pattern forecasting and aiding in early detection of extreme weather events. Despite the crucial data that weather stations provide about sky cloud coverage, their measurements are geographically localized and thus lack spatial coverage. Meteorological satellites on the other hand offer great potential to address this limitation by continuously scanning large areas in short periods of time. This work proposes a novel approach for predicting cloud cover in global satellite images by leveraging ordinal point labels from ground-based weather stations, rather than relying on spatially resolved cloud masks, and demonstrates the effectiveness of this approach using a rank loss-based convolutional neural network of the EfficientNet family. The model is trained in transfer learning approach on a custom-collected dataset across selected regions in the continental USA. Using station measurements only, we achieve an <inline-formula><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula>-score of up to 0.6 and a ranked-within-1-accuracy ranging from 93.5&#x0025; to 99.1&#x0025;. Supplementing the data with labels created by visual inspection to correct for station-satellite mismatches improves scores to 0.75 and 98.4&#x0025; to 100&#x0025;. The results imply significantly improved cloud cover assessment in regions without weather stations, extending the capabilities to monitor localized cloud patterns.
format Article
id doaj-art-181092b73ce346d4b9e4a613b56d730c
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-181092b73ce346d4b9e4a613b56d730c2025-08-20T03:14:01ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118106741068310.1109/JSTARS.2025.355793110949748Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite DataMarkus Gotz0https://orcid.org/0000-0002-2233-1041Erik Wessel1Ugur Cayoglu2https://orcid.org/0000-0002-9670-3717Babak Jahani3https://orcid.org/0000-0002-7347-4878Charlotte Debus4https://orcid.org/0000-0002-7156-2022Jan Cermak5https://orcid.org/0000-0002-4240-595XAchim Streit6https://orcid.org/0000-0002-5065-469XHelmholtz AI, Neuherberg, GermanyScientific Computing Center (SCC), Eggenstein-Leopoldshafen, GermanyScientific Computing Center (SCC), Eggenstein-Leopoldshafen, GermanyInstitute of Meteorology, Climate Research&#x2014;Atmospheric Trace Gases, Remote Sensing, Karlsruhe Institute of Technology, Eggenstein-Leopoldshafen, GermanyScientific Computing Center (SCC), Eggenstein-Leopoldshafen, GermanyInstitute of Meteorology, Climate Research&#x2014;Atmospheric Trace Gases, Remote Sensing, Karlsruhe Institute of Technology, Karlsruhe, GermanyHelmholtz AI, Neuherberg, GermanyAccurate cloud cover assessment is crucial in several fields, such as weather forecasting, climate science, agriculture, or energy system planning, precipitation pattern forecasting and aiding in early detection of extreme weather events. Despite the crucial data that weather stations provide about sky cloud coverage, their measurements are geographically localized and thus lack spatial coverage. Meteorological satellites on the other hand offer great potential to address this limitation by continuously scanning large areas in short periods of time. This work proposes a novel approach for predicting cloud cover in global satellite images by leveraging ordinal point labels from ground-based weather stations, rather than relying on spatially resolved cloud masks, and demonstrates the effectiveness of this approach using a rank loss-based convolutional neural network of the EfficientNet family. The model is trained in transfer learning approach on a custom-collected dataset across selected regions in the continental USA. Using station measurements only, we achieve an <inline-formula><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula>-score of up to 0.6 and a ranked-within-1-accuracy ranging from 93.5&#x0025; to 99.1&#x0025;. Supplementing the data with labels created by visual inspection to correct for station-satellite mismatches improves scores to 0.75 and 98.4&#x0025; to 100&#x0025;. The results imply significantly improved cloud cover assessment in regions without weather stations, extending the capabilities to monitor localized cloud patterns.https://ieeexplore.ieee.org/document/10949748/Convolution neural networksmeteorological aerodrome report (METAR)machine learningordinal classificationsatellite imagessentinel-2
spellingShingle Markus Gotz
Erik Wessel
Ugur Cayoglu
Babak Jahani
Charlotte Debus
Jan Cermak
Achim Streit
Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Convolution neural networks
meteorological aerodrome report (METAR)
machine learning
ordinal classification
satellite images
sentinel-2
title Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
title_full Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
title_fullStr Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
title_full_unstemmed Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
title_short Ordinal METAR Cloud Cover Classification in Sentinel-2 Satellite Data
title_sort ordinal metar cloud cover classification in sentinel 2 satellite data
topic Convolution neural networks
meteorological aerodrome report (METAR)
machine learning
ordinal classification
satellite images
sentinel-2
url https://ieeexplore.ieee.org/document/10949748/
work_keys_str_mv AT markusgotz ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT erikwessel ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT ugurcayoglu ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT babakjahani ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT charlottedebus ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT jancermak ordinalmetarcloudcoverclassificationinsentinel2satellitedata
AT achimstreit ordinalmetarcloudcoverclassificationinsentinel2satellitedata