Articles | Volume 18, issue 11
https://doi.org/10.5194/tc-18-5277-2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/tc-18-5277-2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Pan-Arctic sea ice concentration from SAR and passive microwave
National Center for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
Jørgen Buus-Hinkler
National Center for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
Suman Singha
National Center for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
Hoyeon Shi
National Center for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
Matilde Brandt Kreiner
National Center for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
Related authors
Andreas Stokholm, Jørgen Buus-Hinkler, Tore Wulf, Anton Korosov, Roberto Saldo, Leif Toudal Pedersen, David Arthurs, Ionut Dragan, Iacopo Modica, Juan Pedro, Annekatrien Debien, Xinwei Chen, Muhammed Patel, Fernando Jose Pena Cantu, Javier Noa Turnes, Jinman Park, Linlin Xu, Katharine Andrea Scott, David Anthony Clausi, Yuan Fang, Mingzhe Jiang, Saeid Taleghanidoozdoozan, Neil Curtis Brubacher, Armina Soleymani, Zacharie Gousseau, Michał Smaczny, Patryk Kowalski, Jacek Komorowski, David Rijlaarsdam, Jan Nicolaas van Rijn, Jens Jakobsen, Martin Samuel James Rogers, Nick Hughes, Tom Zagon, Rune Solberg, Nicolas Longépé, and Matilde Brandt Kreiner
The Cryosphere, 18, 3471–3494, https://doi.org/10.5194/tc-18-3471-2024, https://doi.org/10.5194/tc-18-3471-2024, 2024
Short summary
Short summary
The AutoICE challenge encouraged the development of deep learning models to map multiple aspects of sea ice – the amount of sea ice in an area and the age and ice floe size – using multiple sources of satellite and weather data across the Canadian and Greenlandic Arctic. Professionally drawn operational sea ice charts were used as a reference. A total of 179 students and sea ice and AI specialists participated and produced maps in broad agreement with the sea ice charts.
Yi-Jie Yang, Suman Singha, Ron Goldman, and Florian Schütte
Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2025-208, https://doi.org/10.5194/essd-2025-208, 2025
Preprint under review for ESSD
Short summary
Short summary
This data descriptor presents a dataset containing oil slicks, look-alikes, and other remarkable ocean phenomena in synthetic aperture radar (SAR) data, which can be used for training oil spill detection methods. It explains the formation of various oceanic phenomena, supported by examples and supporting materials. These insights can help researchers from diverse backgrounds, such as remote sensing, oceanography, and machine learning, to better understand the sources of the signatures.
Karl Kortum, Suman Singha, and Gunnar Spreen
EGUsphere, https://doi.org/10.5194/egusphere-2024-3351, https://doi.org/10.5194/egusphere-2024-3351, 2024
Short summary
Short summary
Improved sea ice observations are essential to understanding the processes that lead to the strong warming effect currently being observed in the Arctic. In this work, we combine complementary satellite measurement techniques and find remarkable correlations between the two observations. This allows us to expand the coverage of ice topography measurements to a scope and resolution that could not previously be observed.
Guisella Gacitúa, Jacob Lorentsen Høyer, Sten Schmidl Søbjærg, Hoyeon Shi, Sotirios Skarpalezos, Ioanna Karagali, Emy Alerskans, and Craig Donlon
Geosci. Instrum. Method. Data Syst., 13, 373–391, https://doi.org/10.5194/gi-13-373-2024, https://doi.org/10.5194/gi-13-373-2024, 2024
Short summary
Short summary
In spring 2021, a study compared sea surface temperature (SST) measurements from thermal infrared (IR) and passive microwave (PMW) radiometers on a ferry between Denmark and Iceland. The goal was to reduce atmospheric effects and directly compare IR and PMW measurements. A method was developed to convert PMW data to match IR data, with uncertainties analysed in the process. The findings provide insights to improve SST inter-comparisons and enhance the synergy between IR and PMW observations.
Andreas Stokholm, Jørgen Buus-Hinkler, Tore Wulf, Anton Korosov, Roberto Saldo, Leif Toudal Pedersen, David Arthurs, Ionut Dragan, Iacopo Modica, Juan Pedro, Annekatrien Debien, Xinwei Chen, Muhammed Patel, Fernando Jose Pena Cantu, Javier Noa Turnes, Jinman Park, Linlin Xu, Katharine Andrea Scott, David Anthony Clausi, Yuan Fang, Mingzhe Jiang, Saeid Taleghanidoozdoozan, Neil Curtis Brubacher, Armina Soleymani, Zacharie Gousseau, Michał Smaczny, Patryk Kowalski, Jacek Komorowski, David Rijlaarsdam, Jan Nicolaas van Rijn, Jens Jakobsen, Martin Samuel James Rogers, Nick Hughes, Tom Zagon, Rune Solberg, Nicolas Longépé, and Matilde Brandt Kreiner
The Cryosphere, 18, 3471–3494, https://doi.org/10.5194/tc-18-3471-2024, https://doi.org/10.5194/tc-18-3471-2024, 2024
Short summary
Short summary
The AutoICE challenge encouraged the development of deep learning models to map multiple aspects of sea ice – the amount of sea ice in an area and the age and ice floe size – using multiple sources of satellite and weather data across the Canadian and Greenlandic Arctic. Professionally drawn operational sea ice charts were used as a reference. A total of 179 students and sea ice and AI specialists participated and produced maps in broad agreement with the sea ice charts.
Karl Kortum, Suman Singha, Gunnar Spreen, Nils Hutter, Arttu Jutila, and Christian Haas
The Cryosphere, 18, 2207–2222, https://doi.org/10.5194/tc-18-2207-2024, https://doi.org/10.5194/tc-18-2207-2024, 2024
Short summary
Short summary
A dataset of 20 radar satellite acquisitions and near-simultaneous helicopter-based surveys of the ice topography during the MOSAiC expedition is constructed and used to train a variety of deep learning algorithms. The results give realistic insights into the accuracy of retrieval of measured ice classes using modern deep learning models. The models able to learn from the spatial distribution of the measured sea ice classes are shown to have a clear advantage over those that cannot.
Wenkai Guo, Polona Itkin, Suman Singha, Anthony P. Doulgeris, Malin Johansson, and Gunnar Spreen
The Cryosphere, 17, 1279–1297, https://doi.org/10.5194/tc-17-1279-2023, https://doi.org/10.5194/tc-17-1279-2023, 2023
Short summary
Short summary
Sea ice maps are produced to cover the MOSAiC Arctic expedition (2019–2020) and divide sea ice into scientifically meaningful classes. We use a high-resolution X-band synthetic aperture radar dataset and show how image brightness and texture systematically vary across the images. We use an algorithm that reliably corrects this effect and achieve good results, as evaluated by comparisons to ground observations and other studies. The sea ice maps are useful as a basis for future MOSAiC studies.
Thomas Krumpen, Luisa von Albedyll, Helge F. Goessling, Stefan Hendricks, Bennet Juhls, Gunnar Spreen, Sascha Willmes, H. Jakob Belter, Klaus Dethloff, Christian Haas, Lars Kaleschke, Christian Katlein, Xiangshan Tian-Kunze, Robert Ricker, Philip Rostosky, Janna Rückert, Suman Singha, and Julia Sokolova
The Cryosphere, 15, 3897–3920, https://doi.org/10.5194/tc-15-3897-2021, https://doi.org/10.5194/tc-15-3897-2021, 2021
Short summary
Short summary
We use satellite data records collected along the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) drift to categorize ice conditions that shaped and characterized the floe and surroundings during the expedition. A comparison with previous years is made whenever possible. The aim of this analysis is to provide a basis and reference for subsequent research in the six main research areas of atmosphere, ocean, sea ice, biogeochemistry, remote sensing and ecology.
Hoyeon Shi, Byung-Ju Sohn, Gorm Dybkjær, Rasmus Tage Tonboe, and Sang-Moo Lee
The Cryosphere, 14, 3761–3783, https://doi.org/10.5194/tc-14-3761-2020, https://doi.org/10.5194/tc-14-3761-2020, 2020
Short summary
Short summary
To estimate sea ice thickness from satellite freeboard measurements, snow depth information has been required; however, the snow depth estimate has been considered largely uncertain. We propose a new method to estimate sea ice thickness and snow depth simultaneously from freeboards by imposing a thermodynamic constraint. Obtained ice thicknesses and snow depths were consistent with airborne measurements, suggesting that uncertainty of ice thickness caused by uncertain snow depth can be reduced.
Cited articles
Allen, M. J., Dorr, F., Gallego, J. A., Martínez-Ferrer, L., Kalaitzis, F., Ramos-Pollan, R., and Jungbluth, A.: Large Scale Masked Autoencoding for Reducing Label Requirements on SAR Data, in: NeurIPS 2023 Workshop on Tackling Climate Change with Machine Learning, https://www.climatechange.ai/papers/neurips2023/76 (last access: 8 November 2024), 2023. a, b
Allen-Zhu, Z. and Li, Y.: Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning, arXiv [preprint], https://doi.org/10.48550/arXiv.2012.09816, 2020. a
Andersen, S., Tonboe, R., Kaleschke, L., Heygster, G., and Pedersen, L. T.: Intercomparison of passive microwave sea ice concentration retrievals over the high-concentration Arctic sea ice, J. Geophys. Res.-Oceans, 112, C08004, https://doi.org/10.1029/2006JC003543, 2007. a
Asadi, N., Scott, K. A., Komarov, A. S., Buehner, M., and Clausi, D. A.: Evaluation of a Neural Network With Uncertainty for Detection of Ice and Water in SAR Imagery, IEEE T. Geosci. Remote, 59, 247–259, https://doi.org/10.1109/TGRS.2020.2992454, 2021. a
Baordo, F., Vargas, L., and Howe, E.: Algorithm Theoretical Basis Document for Global Sea Ice Concentration Level 2 and Level 3 (OSI-410-a, OSI-401-d, OSI-408-a), https://osisaf-hl.met.no/sites/osisaf-hl/files/baseline_document/osisaf_atbd_ice-conc_l2-3_v1p3.pdf (last access: 8 November 2024), 2023. a
Boulze, H., Korosov, A., and Brajard, J.: Classification of Sea Ice Types in Sentinel-1 SAR Data Using Convolutional Neural Networks, Remote Sens., 12, 13, https://doi.org/10.3390/rs12132165, 2020. a, b
Bourbigot, M., Vincent, P., Johnsen, H., and Piantanida, R.: Sentinel-1 IPF Auxiliary Product Specification, https://sentinels.copernicus.eu/documents/247904/1877131/DI-MPC-PB-0241-3-11-1-Sentinel-1IPFAuxiliaryProductSpecification.pdf/c31f63fa-2db5-d19e-20b0-45e9079d7a04?t=1697806798118 (last access: 8 November 2024), 2023. a
Buus-Hinkler, J., Wulf, T., Stokholm, A. R., Korosov, A., Saldo, R., and Pedersen, L. T.: AI4Arctic/ASIP Sea Ice Dataset – version 2, Technical University of Denmark [data set], https://doi.org/10.11583/DTU.c.6244065.v2, 2022. a, b, c
Caron, M., Touvron, H., Misra, I., Jégou, H., Mairal, J., Bojanowski, P., and Joulin, A.: Emerging Properties in Self-Supervised Vision Transformers, arXiv [preprint], https://doi.org/10.48550/arXiv.2104.14294, 2021. a, b
Caruana, R.: Multitask Learning, Mach. Learn., 28, 41–75, https://doi.org/10.1023/A:1007379606734, 1997. a
Cavalieri, D. J., Gloersen, P., and Campbell, W. J.: Determination of sea ice parameters with the NIMBUS 7 SMMR, J. Geophys. Res.-Atmos., 89, 5355–5369, https://doi.org/10.1029/JD089iD04p05355, 1984. a
Chen, X., Scott, K. A., and Clausi, D. A.: Uncertainty Analysis of Sea Ice and Open Water Classification on SAR Imagery Using a Bayesian CNN, in: Pattern Recognition, Computer Vision, and Image Processing. ICPR 2022 International Workshops and Challenges, edited by: Rousseau, J.-J. and Kapralos, B., Springer Nature Switzerland, Cham, 343–356, ISBN 978-3-031-37731-0, 2023. a
Cheng, A., Casati, B., Tivy, A., Zagon, T., Lemieux, J.-F., and Tremblay, L. B.: Accuracy and inter-analyst agreement of visually estimated sea ice concentrations in Canadian Ice Service ice charts using single-polarization RADARSAT-2, The Cryosphere, 14, 1289–1310, https://doi.org/10.5194/tc-14-1289-2020, 2020. a
Dawkins, H. and Nejadgholi, I.: Region-dependent temperature scaling for certainty calibration and application to class-imbalanced token classification, in: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics, Volume 2: Short Papers, edited by: Muresan, S., Nakov, P., and Villavicencio, A., Association for Computational Linguistics, Dublin, Ireland, 538–544, https://doi.org/10.18653/v1/2022.acl-short.59, 2022. a
EUMETSAT: Global Sea Ice Concentration (AMSR-2), EUMETSAT [data set], https://osi-saf.eumetsat.int/products/osi-408-a (last access: 8 November 2024), 2023. a
European Space Agency: Sentinel-1 SAR User Guide, https://sentiwiki.copernicus.eu/web/s1-products (last access: 8 November 2024), 2023. a
Feng, T., Liu, X., and Li, R.: Super-Resolution-Aided Sea Ice Concentration Estimation From AMSR2 Images by Encoder–Decoder Networks With Atrous Convolution, IEEE J. Sel. To. Appl. Earth Obs., 16, 962–973, https://doi.org/10.1109/JSTARS.2022.3232533, 2023. a
Forster, P., Storelvmo, T., Armour, K., Collins, W., Dufresne, J.-L., Frame, D., Lunt, D., Mauritsen, T., Palmer, M., Watanabe, M., Wild, M., and Zhang, H.: The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 923–1054, https://doi.org/10.1017/9781009157896.009, 2021. a
Frenkel, L. and Goldberger, J.: Network Calibration by Class-based Temperature Scaling, in: 2021 29th European Signal Processing Conference (EUSIPCO), 1486–1490, https://doi.org/10.23919/EUSIPCO54536.2021.9616219, 2021. a
Fuller, A., Millard, K., and Green, J. R.: SatViT: Pretraining Transformers for Earth Observation, IEEE Geosci. Remote Sens., 19, 1–5, https://doi.org/10.1109/LGRS.2022.3201489, 2022. a
He, K., Chen, X., Xie, S., Li, Y., Dollár, P., and Girshick, R. B.: Masked Autoencoders Are Scalable Vision Learners, arXiv [preprint], https://doi.org/10.48550/arXiv.2111.06377, 2021. a, b, c
Hendrycks, D. and Gimpel, K.: Bridging Nonlinearities and Stochastic Regularizers with Gaussian Error Linear Units, arXiv [preprint], https://doi.org/10.48550/arXiv.1606.08415, 2016. a
Huang, G., Sun, Y., Liu, Z., Sedra, D., and Weinberger, K.: Deep Networks with Stochastic Depth, Computer Vision – ECCV 2016, Springer International Publishing, 646–661, https://doi.org/10.1007/978-3-319-46493-0_39, 2016. a
International Ice Charting Working Group: IICWG Task Team 4 & 12 on Uncertainty, https://nsidc.org/noaa/iicwg/task-teams (last access: 8 November 2024), 2021. a
International Maritime Organization, Maritime Safety Committee: Guidance on Methodologies for Assessing Operational Capabilities and Limitations in Ice – Appendix - Methodology for Assessing Operational Capabilities and Limitations in Ice: Polar Operational Limit Assessment Risk Indexing System (POLARIS), https://www.imorules.com/GUID-2C1D86CB-5D58-490F-B4D4-46C057E1D102.html (last access: 8 November 2024), 2016. a
Ioffe, S. and Szegedy, C.: Batch normalization: accelerating deep network training by reducing internal covariate shift, ICML'15, JMLR.org, 448–456, https://doi.org/10.5555/3045118.3045167, 2015. a
Iris, D. G., Colin, A., and Longépé, N.: Prediction of Categorized Sea Ice Concentration From Sentinel-1 SAR Images Based on a Fully Convolutional Network, IEEE J. Sel. Top. Appl. Earth Obs., 14, 5831–5841, https://doi.org/10.1109/JSTARS.2021.3074068, 2021. a
Jakubik, J., Roy, S., Phillips, C. E., Fraccaro, P., Godwin, D., Zadrozny, B., Szwarcman, D., Gomes, C., Nyirjesy, G., Edwards, B., Kimura, D., Simumba, N., Chu, L., Mukkavilli, S. K., Lambhate, D., Das, K., Bangalore, R., Oliveira, D., Muszynski, M., Ankur, K., Ramasubramanian, M., Gurung, I., Khallaghi, S., Hanxi, Li, Cecil, M., Ahmadi, M., Kordi, F., Alemohammad, H., Maskey, M., Ganti, R., Weldemariam, K., and Ramachandran, R.: Foundation Models for Generalist Geospatial Artificial Intelligence, arXiv [preprint], https://doi.org/10.48550/arXiv.2310.18660, 2023. a
Karvonen, J.: Baltic Sea ice SAR segmentation and classification using modified pulse-coupled neural networks, IEEE T. Geosci. Remote, 42, 1566–1574, https://doi.org/10.1109/TGRS.2004.828179, 2004. a
Karvonen, J.: Baltic Sea Ice Concentration Estimation Using SENTINEL-1 SAR and AMSR2 Microwave Radiometer Data, IEEE T. Geosci. Remote, 55, 2871–2883, https://doi.org/10.1109/TGRS.2017.2655567, 2017. a
Karvonen, J.: Baltic Sea Ice Concentration Estimation From C-Band Dual-Polarized SAR Imagery by Image Segmentation and Convolutional Neural Networks, IEEE T. Geosci. Remote, 60, 1–11, https://doi.org/10.1109/TGRS.2021.3097885, 2022. a
Karvonen, J., Vainio, J., Marnela, M., Eriksson, P., and Niskanen, T.: A Comparison Between High-Resolution EO-Based and Ice Analyst-Assigned Sea Ice Concentrations, IEEE J. Sel. Top. Appl. Earth Obs., 8, 1799–1807, https://doi.org/10.1109/JSTARS.2015.2426414, 2015. a, b
Khaleghian, S., Ullah, H., Kræmer, T., Hughes, N., Eltoft, T., and Marinoni, A.: Sea Ice Classification of SAR Imagery Based on Convolution Neural Networks, Remote Sens., 13, 1734, https://doi.org/10.3390/rs13091734, 2021. a, b
Korosov, A., Johansson, M., Shuchman, R., Dierking, W., and Kwok, R.: SeaSAR Theme 4: Sea Ice, Sea Ice Drift, Sea Ice Type, Multi-sensor synergy, In situ campaigns, https://eo4society.esa.int/training_uploads/seasar2023/Seasar2023_Theme_4_Sea_Ice_Retrievals_part1.pdf (last access: 8 November 2024), 2023. a
Kortum, K., Singha, S., and Spreen, G.: Robust Multiseasonal Ice Classification From High-Resolution X-Band SAR, IEEE T. Geosci. Remote, 60, 1–12, https://doi.org/10.1109/TGRS.2022.3144731, 2022. a, b
Kortum, K., Singha, S., Spreen, G., Hutter, N., Jutila, A., and Haas, C.: SAR deep learning sea ice retrieval trained with airborne laser scanner measurements from the MOSAiC expedition, The Cryosphere, 18, 2207–2222, https://doi.org/10.5194/tc-18-2207-2024, 2024. a
Kreiner, M. B., Wulf, T., Jakobsen, J., Nielsen, A. A., and Pedersen, L. T.: Inter- and intra-analyst ice edge assessment, figshare [data set], https://doi.org/10.6084/m9.figshare.22312648.v1, 2023. a
Kull, M., Perello-Nieto, M., Kängsepp, M., Filho, T. S., Song, H., and Flach, P.: Beyond temperature scaling: Obtaining well-calibrated multiclass probabilities with Dirichlet calibration, Proceedings of the 33rd International Conference on Neural Information Processing Systems, 1103, 11 pp., 2019. a, b
Lakshminarayanan, B., Pritzel, A., and Blundell, C.: Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles, in: Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS'17, Curran Associates Inc., 6405–6416, ISBN 9781510860964, 2017. a, b, c, d, e, f, g
Lavergne, T., Sørensen, A. M., Kern, S., Tonboe, R., Notz, D., Aaboe, S., Bell, L., Dybkjær, G., Eastwood, S., Gabarro, C., Heygster, G., Killie, M. A., Brandt Kreiner, M., Lavelle, J., Saldo, R., Sandven, S., and Pedersen, L. T.: Version 2 of the EUMETSAT OSI SAF and ESA CCI sea-ice concentration climate data records, The Cryosphere, 13, 49–78, https://doi.org/10.5194/tc-13-49-2019, 2019. a
Liu, Z., Mao, H., Wu, C., Feichtenhofer, C., Darrell, T., and Xie, S.: A ConvNet for the 2020s, arXiv [preprint], https://doi.org/10.48550/arXiv.2201.03545, 2022. a
Loshchilov, I. and Hutter, F.: Fixing Weight Decay Regularization in Adam, arXiv [preprint], https://doi.org/10.48550/arXiv.1711.05101, 2017. a
Malmgren-Hansen, D., Pedersen, L. T., Nielsen, A. A., Skriver, H., Saldo, R., Kreiner, M. B., and Buus-Hinkler, J.: ASIP Sea Ice Dataset – version 1, DTU Data [data set], https://doi.org/10.11583/DTU.11920416.v1, 2020. a, b
Malmgren-Hansen, D., Pedersen, L. T., Nielsen, A. A., Kreiner, M. B., Saldo, R., Skriver, H., Lavelle, J., Buus-Hinkler, J., and Krane, K. H.: A Convolutional Neural Network Architecture for Sentinel-1 and AMSR2 Data Fusion, IEEE T. Geosci. Remote, 59, 1890–1902, https://doi.org/10.1109/TGRS.2020.3004539, 2021. a, b, c, d, e
Meier, W. N. and Stewart, J. S.: Assessing the Potential of Enhanced Resolution Gridded Passive Microwave Brightness Temperatures for Retrieval of Sea Ice Parameters, Remote Sens., 12, 16, https://doi.org/10.3390/rs12162552, 2020. a
Moon, T. A., Thoman, R., Druckenmiller, M. L., Ahmasuk, B., Backensto, S. A., Ballinger, T. J., Benestad, R., Berner, L. T., Bernhard, G. H., Bhatt, U. S., Bigalke, S., BjerkeJarle, W., Brettschneider, B., Christiansen, H. H., Cohen, J. L., Decharme, B., Derksen, C., Divine, D., Drost, J., Druckenmiller, M. L., EliasChereque, A., Epstein, H. E., Fausto, R. S., Fettweis, X., Fioletov, V. E., Forbes, B. C., Frost, G. V., , Gerland, S., Goetz, S. J., Grooß, J.-U., Hanna, E., Hanssen-Bauer, I., Hendricks, S., Holmes, R. M., Ialongo, I., Isaksen, K., Johnsen, B., Jones, T., Kaler, R. S., Kaleschke, L., Kim, S.-J., Labe, Z. M., Lader, R., Lakkala, K., Lara, M. J., Lindsey, J., Loomis, B. D., Luojus, K., Macander, M. J., Mamen, J., Mankoff, K. D., Manney, G. L., McAfee, S. A., McClelland, J. W., Meier, W. N., Moon, T. A., Moore, G. W. K., Mote, T. L., Mudryk, L., Müller, R., Nyland, K. E., Overland, J. E., Parrish, J. K., Perovich, D. K., Petersen, G. N., Petty, A., Phoenix, G. K., Poinar, K., Rantanen, M., Ricker, R., Romanovsky, V. E., Serbin, S. P., Serreze, M. C., Sheffield, G., Shiklomanov, A. I., Shiklomanov, N. I., Smith, S. L., Spencer, R. G. M., Streletskiy, D. A., Suslova, A., Svendby, T., Tank, S. E., Tedesco, M., Thoman, R. L., Tian-Kunze, X., Timmermans, M.-L., Tømmervik, H., Tretiakov, M., Walker, D. A., Walsh, J. E., Wang, M., Webster, M., Wehrlé, A., Yang, D., Zolkos, S., Allen, J., Camper, A. V., Haley, B. O., Hammer, G., Love-Brotak, S., Ohlmann, L., Noguchi, L., Riddle, D. B., and Veasey, S. W.: The Arctic, B. Am. Meteorol. Soc., 104, S271–S321, https://doi.org/10.1175/BAMS-D-23-0079.1, 2023. a
Murphy, A. H. and Winkler, R. L.: Reliability of Subjective Probability Forecasts of Precipitation and Temperature, J. Roy. Stat. Soc. Ser. C, 26, 41–47, https://doi.org/10.2307/2346866, 1977. a
Naeini, M. P., Cooper, G. F., and Hauskrecht, M.: Obtaining Well Calibrated Probabilities Using Bayesian Binning, in: AAAI, 2901–2907, https://doi.org/10.1609/aaai.v29i1.9602, 2015. a
Ovadia, Y., Fertig, E., Ren, J., Nado, Z., Sculley, D., Nowozin, S., Dillon, J. V., Lakshminarayanan, B., and Snoek, J.: Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift, in: Proceedings of the 33rd International Conference on Neural Information Processing Systems, Curran Associates Inc., 1254, 12 pp., 2019. a, b, c
Pires de Lima, R. and Karimzadeh, M.: Model Ensemble with Dropout for Uncertainty Estimation in Binary Sea Ice or Water Segmentation using Sentinel-1 SAR, IEEE T. Geosci. Remote, 61, 1–15, https://doi.org/10.1109/TGRS.2023.3331276, 2023. a
Ponsoni, L., Ribergaard, M. H., Nielsen-Englyst, P., Wulf, T., Buus-Hinkler, J., Kreiner, M. B., and Rasmussen, T. A. S.: Greenlandic sea ice products with a focus on an updated operational forecast system, Front. Mar. Sci., 10, 979782, https://doi.org/10.3389/fmars.2023.979782, 2023. a
Ptresample developers: Pyresample, Ptresample developers [code], https://pyresample.readthedocs.io/en/latest (last access: 8 November 2024), 2023. a
Ressel, R., Singha, S., Lehner, S., Rösel, A., and Spreen, G.: Investigation into Different Polarimetric Features for Sea Ice Classification Using X-Band Synthetic Aperture Radar, IEEE J. Sel. Top. Appl. Earth Obs., 9, 3131–3143, https://doi.org/10.1109/JSTARS.2016.2539501, 2016. a
Ronneberger, O., Fischer, P., and Brox, T.: U-Net: Convolutional Networks for Biomedical Image Segmentation, in: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, Springer International Publishing, 234–241, https://doi.org/10.1007/978-3-319-24574-4_28, 2015. a
Saldo, R., Brandt Kreiner, M., Buus-Hinkler, J., Pedersen, L. T., Malmgren-Hansen, D., and Nielsen, A. A.: AI4Arctic/ASIP Sea Ice Dataset – version 2, DTU Data [data set], https://doi.org/10.11583/DTU.13011134.v2, 2020. a, b, c
Sandler, M., Howard, A. G., Zhu, M., Zhmoginov, A., and Chen, L.: Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segmentation, arXiv [preprint], https://doi.org/10.48550/arXiv.1801.04381, 2018. a, b
Singha, S., Johansson, A. M., Hughes, N., Munk Hvidegaard, S., and Skourup, H.: Arctic Sea Ice Characterization using Spaceborne Fully Polarimetric L-, C- and X-band SAR with Validation by Airborne Measurements, IEEE T. Geosci. Remote, 56, 3715–3734, https://doi.org/10.1109/TGRS.2018.2809504, 2018. a
Stokholm, A., Wulf, T., Kucik, A., Saldo, R., Buus-Hinkler, J., and Hvidegaard, S. M.: AI4SeaIce: Toward Solving Ambiguous SAR Textures in Convolutional Neural Networks for Automatic Sea Ice Concentration Charting, IEEE T. Geosci. Remote, 60, 1–13, https://doi.org/10.1109/TGRS.2022.3149323, 2022. a, b, c, d, e
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., and Wojna, Z.: Rethinking the Inception Architecture for Computer Vision, arXiv [preprint], https://doi.org/10.48550/arXiv.1512.00567, 2015. a, b
Thulasidasan, S., Chennupati, G., Bilmes, J., Bhattacharya, T., and Michalak, S.: On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks, ArXiv [preprint], https://doi.org/10.48550/arXiv.1905.11001, 2020. a, b
Tonboe, R. T., Eastwood, S., Lavergne, T., Sørensen, A. M., Rathmann, N., Dybkjær, G., Pedersen, L. T., Høyer, J. L., and Kern, S.: The EUMETSAT sea ice concentration climate data record, The Cryosphere, 10, 2275–2290, https://doi.org/10.5194/tc-10-2275-2016, 2016. a
Touvron, H., Cord, M., Sablayrolles, A., Synnaeve, G., and Jégou, H.: Going deeper with Image Transformers, 2021 IEEE/CVF International Conference on Computer Vision (ICCV), IEEE Computer Society, Los Alamitos, CA, USA, 32–42, https://doi.org/10.1109/ICCV48922.2021.00010, 2021. a
U.S. National Ice Center: U.S. National Ice Center Arctic and Antarctic Sea Ice Charts in SIGRID-3 Format, Version 1, National Snow and Ice Data Center [data set], https://doi.org/10.7265/4b7s-rn93, 2022. a
Wang, L., Scott, K. A., Xu, L., and Clausi, D. A.: Sea Ice Concentration Estimation During Melt From Dual-Pol SAR Scenes Using Deep Convolutional Neural Networks: A Case Study, IEEE T. Geosci. Remote, 54, 4524–4533, https://doi.org/10.1109/TGRS.2016.2543660, 2016. a
Wang, Y.-R. and Li, X.-M.: Arctic sea ice cover data from spaceborne synthetic aperture radar by deep learning, Earth Syst. Sci. Data, 13, 2723–2742, https://doi.org/10.5194/essd-13-2723-2021, 2021. a
Wen, Y., Tran, D., and Ba, J.: BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning, arXiv [preprint], https://doi.org/10.48550/arXiv.2002.06715, 2020. a
Wulf, T., Buus-Hinkler, J., and Kreiner, M. B.: Synergetic Fusion of Satellite SAR and Passive Microwave Radiometer Observations for Automatic Sea Ice Charting using Convolutional Neural Networks, in: ESA Living Planet Symposium, ESA, Bonn, Germany, https://lps22-programme.esa.int/posters/POSTERS_Day5.pdf (last access: 8 November 2024), 2022. a, b
Wulf, T., Buus-Hinkler, J., Singha, S., and Kreiner, M. B.: Operational high resolution Arctic sea ice concentration retrieval using SAR and passive microwave observations, in: nternational Symposium on Sea Ice Across Temporal and Spatial Scales, IGS, Bremerhaven, Germany, https://www.igsoc.org/wp-content/uploads/2023/06/programme_80.html (last access: 8 November 2024), 2023a. a
Wulf, T., Buus-Hinkler, J., Singha, S., and Kreiner, M. B.: Fusion of satellite SAR and passive microwave radiometer observations for automatic sea ice mapping using convolutional neural networks, in: Book of abstracts for the IICWG-DA-11 workshop, IICWG, Oslo, Norway, https://iicwg-da-11.met.no/programme (last access: 8 November 2024), 2023b. a
Zakhvatkina, N. Y., Alexandrov, V. Y., Johannessen, O. M., Sandven, S., and Frolov, I. Y.: Classification of Sea Ice Types in ENVISAT Synthetic Aperture Radar Images, IEEE T. Geosci. Remote, 51, 2587–2600, https://doi.org/10.1109/TGRS.2012.2212445, 2013. a
Short summary
Here, we present ASIP: a new and comprehensive deep-learning-based methodology to retrieve high-resolution sea ice concentration with accompanying well-calibrated uncertainties from satellite-based active and passive microwave observations at a pan-Arctic scale for all seasons. In a comparative study against pan-Arctic ice charts and well-established passive-microwave-based sea ice products, we show that ASIP generalizes well to the pan-Arctic region.
Here, we present ASIP: a new and comprehensive deep-learning-based methodology to retrieve...