{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,25]],"date-time":"2026-09-25T00:23:14Z","timestamp":1790295794572,"version":"4.1.0"},"reference-count":73,"publisher":"American Geophysical Union (AGU)","issue":"12","license":[{"start":{"date-parts":[[2022,12,14]],"date-time":"2022-12-14T00:00:00Z","timestamp":1670976000000},"content-version":"vor","delay-in-days":13,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"DOI":"10.13039\/100010663","name":"European Research Council","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100010663","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["agupubs.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["J Adv Model Earth Syst"],"published-print":{"date-parts":[[2022,12]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>A promising approach to improve cloud parameterizations within climate models and thus climate projections is to use deep learning in combination with training data from storm\u2010resolving model (SRM) simulations. The ICOsahedral Non\u2010hydrostatic (ICON) modeling framework permits simulations ranging from numerical weather prediction to climate projections, making it an ideal target to develop neural network (NN) based parameterizations for sub\u2010grid scale processes. Within the ICON framework, we train NN based cloud cover parameterizations with coarse\u2010grained data based on realistic regional and global ICON SRM simulations. We set up three different types of NNs that differ in the degree of vertical locality they assume for diagnosing cloud cover from coarse\u2010grained atmospheric state variables. The NNs accurately estimate sub\u2010grid scale cloud cover from coarse\u2010grained data that has similar geographical characteristics as their training data. Additionally, globally trained NNs can reproduce sub\u2010grid scale cloud cover of the regional SRM simulation. Using the game\u2010theory based interpretability library SHapley Additive exPlanations, we identify an overemphasis on specific humidity and cloud ice as the reason why our column\u2010based NN cannot perfectly generalize from the global to the regional coarse\u2010grained SRM data. The interpretability tool also helps visualize similarities and differences in feature importance between regionally and globally trained column\u2010based NNs, and reveals a local relationship between their cloud cover predictions and the thermodynamic environment. Our results show the potential of deep learning to derive accurate yet interpretable cloud cover parameterizations from global SRMs, and suggest that neighborhood\u2010based models may be a good compromise between accuracy and generalizability.<\/jats:p>","DOI":"10.1029\/2021ms002959","type":"journal-article","created":{"date-parts":[[2022,12,8]],"date-time":"2022-12-08T20:33:42Z","timestamp":1670531622000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":44,"title":["Deep Learning Based Cloud Cover Parameterization for ICON"],"prefix":"10.1029","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3765-242X","authenticated-orcid":false,"given":"Arthur","family":"Grundner","sequence":"first","affiliation":[{"name":"Deutsches Zentrum f\u00fcr Luft\u2010 und Raumfahrt e.V. (DLR) Institut f\u00fcr Physik der Atmosph\u00e4re  Oberpfaffenhofen Germany"},{"name":"Center for Learning the Earth with Artificial Intelligence and Physics (LEAP) Columbia University  New York NY USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5731-1040","authenticated-orcid":false,"given":"Tom","family":"Beucler","sequence":"additional","affiliation":[{"name":"Institute of Earth Surface Dynamics University of Lausanne  Lausanne Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0845-8345","authenticated-orcid":false,"given":"Pierre","family":"Gentine","sequence":"additional","affiliation":[{"name":"Center for Learning the Earth with Artificial Intelligence and Physics (LEAP) Columbia University  New York NY USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3403-8245","authenticated-orcid":false,"given":"Fernando","family":"Iglesias\u2010Suarez","sequence":"additional","affiliation":[{"name":"Deutsches Zentrum f\u00fcr Luft\u2010 und Raumfahrt e.V. (DLR) Institut f\u00fcr Physik der Atmosph\u00e4re  Oberpfaffenhofen Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4278-1963","authenticated-orcid":false,"given":"Marco A.","family":"Giorgetta","sequence":"additional","affiliation":[{"name":"Max Planck Institute for Meteorology  Hamburg Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6887-4885","authenticated-orcid":false,"given":"Veronika","family":"Eyring","sequence":"additional","affiliation":[{"name":"Deutsches Zentrum f\u00fcr Luft\u2010 und Raumfahrt e.V. (DLR) Institut f\u00fcr Physik der Atmosph\u00e4re  Oberpfaffenhofen Germany"},{"name":"University of Bremen Institute of Environmental Physics (IUP)  Bremen Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"13","published-online":{"date-parts":[[2022,12,14]]},"reference":[{"key":"e_1_2_10_2_1","doi-asserted-by":"publisher","DOI":"10.1038\/nature01092"},{"key":"e_1_2_10_3_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-28954-6_9"},{"key":"e_1_2_10_4_1","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0442(2003)016<2676:adasrt>2.0.co;2"},{"key":"e_1_2_10_5_1","doi-asserted-by":"crossref","unstructured":"Beucler T. Pritchard M. Gentine P. &Rasp S.(2020).Towards physically\u2010consistent data\u2010driven models of convection. InIgarss 2020\u20102020 ieee international geoscience and remote sensing symposium.https:\/\/doi.org\/10.1109\/IGARSS39084.2020.9324569","DOI":"10.1109\/IGARSS39084.2020.9324569"},{"key":"e_1_2_10_6_1","unstructured":"Beucler T. Pritchard M. Yuval J. Gupta A. Peng L. Rasp S. et\u00a0al. (2021).Climate\u2010invariant machine learning. arXiv preprint arXiv:2112.08440."},{"key":"e_1_2_10_7_1","doi-asserted-by":"publisher","DOI":"10.1175\/JAS-D-20-0082.1"},{"key":"e_1_2_10_8_1","doi-asserted-by":"publisher","DOI":"10.1029\/2018gl078510"},{"key":"e_1_2_10_9_1","doi-asserted-by":"publisher","DOI":"10.1029\/2019ms001711"},{"key":"e_1_2_10_10_1","doi-asserted-by":"publisher","DOI":"10.1175\/JAS3467.1"},{"key":"e_1_2_10_11_1","doi-asserted-by":"publisher","DOI":"10.1029\/2021MS002477"},{"key":"e_1_2_10_12_1","doi-asserted-by":"publisher","DOI":"10.1002\/qj.49712656318"},{"key":"e_1_2_10_13_1","doi-asserted-by":"publisher","DOI":"10.1029\/2017ms001233"},{"key":"e_1_2_10_14_1","volume-title":"A description of the nonhydrostatic regional Cosmo model, part II: Physical parameterization","author":"Doms G.","year":"2011"},{"key":"e_1_2_10_15_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41558-021-01020-x"},{"key":"e_1_2_10_16_1","doi-asserted-by":"publisher","DOI":"10.1029\/2019ms001896"},{"key":"e_1_2_10_17_1","unstructured":"Gal Y. &Ghahramani Z.(2016).Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. InProceedings of the 33rd international conference on machine learning."},{"key":"e_1_2_10_18_1","doi-asserted-by":"crossref","first-page":"307","DOI":"10.1002\/9781119646181.ch21","article-title":"Deep learning for the parametrization of subgrid processes in climate models","author":"Gentine P.","year":"2021","journal-title":"Deep Learning for the Earth Sciences: A Comprehensive Approach to Remote Sensing, Climate Science, and Geosciences"},{"key":"e_1_2_10_19_1","doi-asserted-by":"publisher","DOI":"10.1029\/2018gl078202"},{"key":"e_1_2_10_20_1","doi-asserted-by":"publisher","DOI":"10.1029\/2020ms002268"},{"key":"e_1_2_10_21_1","doi-asserted-by":"publisher","DOI":"10.1029\/2017ms001233"},{"key":"e_1_2_10_22_1","doi-asserted-by":"publisher","DOI":"10.5194\/gmd-15-6985-2022"},{"key":"e_1_2_10_23_1","doi-asserted-by":"publisher","DOI":"10.1029\/2020ms002076"},{"key":"e_1_2_10_24_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.softx.2020.100591"},{"key":"e_1_2_10_25_1","doi-asserted-by":"publisher","DOI":"10.2151\/jmsj.2020-005"},{"key":"e_1_2_10_26_1","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(91)90009-t"},{"key":"e_1_2_10_27_1","doi-asserted-by":"publisher","DOI":"10.1175\/JAS3453.1"},{"key":"e_1_2_10_28_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41561-017-0005-4"},{"key":"e_1_2_10_29_1","doi-asserted-by":"publisher","DOI":"10.1155\/2013\/485913"},{"key":"e_1_2_10_30_1","doi-asserted-by":"publisher","DOI":"10.1175\/MWR2923.1"},{"key":"e_1_2_10_31_1","unstructured":"Larson V. E.(2017).Clubb\u2010silhs: A parameterization of subgrid variability in the atmosphere. arXiv preprint arXiv:1711.03675."},{"key":"e_1_2_10_32_1","doi-asserted-by":"publisher","DOI":"10.1007\/BF00207939"},{"key":"e_1_2_10_33_1","unstructured":"Lundberg S. M. &Lee S.\u2010I.(2017).A unified approach to interpreting model predictions. In31st conference on neural information processing systems."},{"key":"e_1_2_10_34_1","doi-asserted-by":"publisher","DOI":"10.1029\/2018ms001400"},{"key":"e_1_2_10_35_1","doi-asserted-by":"publisher","DOI":"10.1175\/2007jas2294.1"},{"key":"e_1_2_10_36_1","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.aba1981"},{"key":"e_1_2_10_37_1","doi-asserted-by":"publisher","DOI":"10.1029\/97jd00237"},{"key":"e_1_2_10_38_1","volume-title":"Interpretable machine learning","author":"Molnar C.","year":"2020"},{"key":"e_1_2_10_39_1","doi-asserted-by":"crossref","unstructured":"Mooers G. Tuyls J. Mandt S. Pritchard M. &Beucler T.(2020).Generative modeling of atmospheric convection.Proceedings of the 10th international conference on climate informatics.https:\/\/doi.org\/10.1145\/3429309.3429324","DOI":"10.1145\/3429309.3429324"},{"key":"e_1_2_10_40_1","volume-title":"Convectively generated gravity waves and convective aggregation in numerical models of tropical dynamics","author":"M\u00fcller S.","year":"2019"},{"key":"e_1_2_10_41_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-020-15195-y"},{"key":"e_1_2_10_42_1","first-page":"1","volume-title":"A fortran\u2010keras deep learning bridge for scientific computing","author":"Ott J.","year":"2020"},{"key":"e_1_2_10_43_1","doi-asserted-by":"publisher","DOI":"10.1029\/2019MS001621"},{"key":"e_1_2_10_44_1","doi-asserted-by":"publisher","DOI":"10.1002\/jame.20027"},{"key":"e_1_2_10_45_1","unstructured":"Prill F. Reinert D. Rieger D. Z\u00e4ngl G. Schr\u00f6ter J. F\u00f6rstner J. et\u00a0al. (2019).Icon tutorial: Nwp mode and icon\u2010art[Computer software manual].ICON. Retrieved fromhttps:\/\/code.mpimet.mpg.de\/attachments\/19568\/ICON_tutorial_2019.pdf"},{"key":"e_1_2_10_46_1","doi-asserted-by":"publisher","DOI":"10.1007\/s00382-007-0247-8"},{"key":"e_1_2_10_47_1","doi-asserted-by":"publisher","DOI":"10.1175\/bams-84-11-1547"},{"key":"e_1_2_10_48_1","first-page":"89","article-title":"The new turbulence parameterization of lm","volume":"1","author":"Raschendorfer M.","year":"2001","journal-title":"COSMO newsletter"},{"key":"e_1_2_10_49_1","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1810286115"},{"key":"e_1_2_10_50_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-019-10105-3"},{"key":"e_1_2_10_51_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-28954-6"},{"key":"e_1_2_10_52_1","doi-asserted-by":"publisher","DOI":"10.5194\/esd-11-1233-2020"},{"key":"e_1_2_10_53_1","doi-asserted-by":"publisher","DOI":"10.1038\/nclimate3190"},{"key":"e_1_2_10_54_1","volume-title":"The multi\u2010layer version of the dwd soil model terra_lm","author":"Schrodin R.","year":"2001"},{"key":"e_1_2_10_55_1","doi-asserted-by":"crossref","unstructured":"Schulz J.\u2010P. Vogel G. Becker C. Kothe S. &Ahrens B.(2015).Evaluation of the ground heat flux simulated by a multi\u2010layer land surface scheme using high\u2010quality observations at grass land and bare soil. InEgu general assembly conference abstracts(p.6549).","DOI":"10.1127\/metz\/2016\/0537"},{"key":"e_1_2_10_56_1","unstructured":"Schulzweida U.(2019).Cdo user guide.https:\/\/doi.org\/10.5281\/zenodo.3539275"},{"key":"e_1_2_10_57_1","first-page":"25","article-title":"A revised cloud microphysical parameterization for COSMO\u2010LME","volume":"7","author":"Seifert A.","year":"2008","journal-title":"COSMO Newsletter"},{"key":"e_1_2_10_58_1","doi-asserted-by":"publisher","DOI":"10.1029\/2020ms002301"},{"key":"e_1_2_10_59_1","unstructured":"Shrikumar A. Greenside P. &Kundaje A.(2017).Learning important features through propagating activation differences. InInternational conference on machine learning(pp.3145\u20133153)."},{"key":"e_1_2_10_60_1","volume-title":"Parameterization schemes: Keys to understanding numerical weather prediction models","author":"Stensrud D. J.","year":"2009"},{"key":"e_1_2_10_61_1","doi-asserted-by":"publisher","DOI":"10.2151\/jmsj.2020-021"},{"key":"e_1_2_10_62_1","doi-asserted-by":"publisher","DOI":"10.1175\/bams-d-18-0198.1"},{"key":"e_1_2_10_63_1","doi-asserted-by":"publisher","DOI":"10.1186\/s40645-019-0304-z"},{"key":"e_1_2_10_64_1","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0493(1989)117<1641:cacpsw>2.0.co;2"},{"key":"e_1_2_10_65_1","doi-asserted-by":"publisher","DOI":"10.1006\/jcph.2001.6897"},{"key":"e_1_2_10_66_1","volume-title":"ECMWF moist processes lecture note series tech","author":"Tompkins A.","year":"2005"},{"key":"e_1_2_10_67_1","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0469(2002)059\u23291917:APPFTS\u232a2.0.CO;2"},{"key":"e_1_2_10_68_1","doi-asserted-by":"publisher","DOI":"10.1175\/jcli-d-19-0286.1"},{"key":"e_1_2_10_69_1","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0493(1994)122\u23291021:CCAIRT\u232a2.0.CO;2"},{"key":"e_1_2_10_70_1","doi-asserted-by":"publisher","DOI":"10.5194\/gmd-15-3923-2022"},{"key":"e_1_2_10_71_1","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0493(1991)119\u23290342:EOCPUA\u232a2.0.CO;2"},{"key":"e_1_2_10_72_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-020-17142-3"},{"key":"e_1_2_10_73_1","doi-asserted-by":"publisher","DOI":"10.1029\/2020gl091363"},{"key":"e_1_2_10_74_1","doi-asserted-by":"publisher","DOI":"10.1002\/qj.2378"}],"container-title":["Journal of Advances in Modeling Earth Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1029\/2021MS002959","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1029\/2021MS002959","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/agupubs.onlinelibrary.wiley.com\/doi\/pdf\/10.1029\/2021MS002959","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,20]],"date-time":"2023-08-20T09:17:45Z","timestamp":1692523065000},"score":1,"resource":{"primary":{"URL":"https:\/\/agupubs.onlinelibrary.wiley.com\/doi\/10.1029\/2021MS002959"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,12]]},"references-count":73,"journal-issue":{"issue":"12","published-print":{"date-parts":[[2022,12]]}},"alternative-id":["10.1029\/2021MS002959"],"URL":"https:\/\/doi.org\/10.1029\/2021ms002959","archive":["Portico"],"relation":{"has-preprint":[{"id-type":"doi","id":"10.1002\/essoar.10509777.1","asserted-by":"object"}]},"ISSN":["1942-2466","1942-2466"],"issn-type":[{"value":"1942-2466","type":"print"},{"value":"1942-2466","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,12]]},"assertion":[{"value":"2021-12-20","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-12-05","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2022-12-14","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e2021MS002959"}}