{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T09:14:50Z","timestamp":1766049290845,"version":"3.48.0"},"reference-count":43,"publisher":"IOP Publishing","issue":"4","license":[{"start":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T00:00:00Z","timestamp":1766016000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T00:00:00Z","timestamp":1766016000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"DOI":"10.13039\/100007000","name":"Laboratory Directed Research and Development","doi-asserted-by":"crossref","award":["Project No. 22-SI-008"],"award-info":[{"award-number":["Project No. 22-SI-008"]}],"id":[{"id":"10.13039\/100007000","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2025,12,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Deep learning (DL) has emerged as a promising tool for downscaling coarse-resolution climate data to high-resolution outputs, enabling improved regional climate predictions. A critical aspect of DL-based downscaling is the incorporation of uncertainty quantification (UQ), which enhances the interpretability and reliability of predictions\u2014key factors for climate risk assessment and decision-making. This study develops a DL model to downscale 2 m temperature across the contiguous United States using reanalysis datasets. We systematically evaluate three epistemic UQ methods\u2014deep ensembles (DEns), Monte Carlo dropout (MCD), and Flipout\u2014based on their probabilistic accuracy, downscaling performance, sensitivity to geographical features, and computational efficiency. Results indicate that MCD generally outperforms Flipout and DEns in terms of calibration and downscaling accuracy. However, DEns demonstrate lower calibration errors in coastal regions, indicating its higher confidence within these areas. Flipout, in contrast, is more sensitive to elevation gradients and exhibits higher calibration errors in mountainous regions. Hence, the choice of UQ method for this task depends on the specific requirements of the application. For applications that prioritize overall calibration, downscaling accuracy, and computational efficiency, MCD is a strong candidate. These findings highlight the importance of selecting UQ methods based on application-specific requirements, such as geographical context and computational constraints. By addressing the trade-offs between UQ methods, this study provides actionable insights for improving the reliability, scalability, and utility of DL-based downscaling in climate science.<\/jats:p>","DOI":"10.1088\/2632-2153\/ae281e","type":"journal-article","created":{"date-parts":[[2025,12,4]],"date-time":"2025-12-04T22:54:52Z","timestamp":1764888892000},"page":"045069","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Evaluating probabilistic deep learning methods for uncertainty quantification of temperature downscaling"],"prefix":"10.1088","volume":"6","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2594-7845","authenticated-orcid":true,"given":"Yannic","family":"Lops","sequence":"first","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4169-5553","authenticated-orcid":true,"given":"Gemma J","family":"Anderson","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4649-6967","authenticated-orcid":true,"given":"Donald D","family":"Lucas","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0132-8417","authenticated-orcid":false,"given":"Indrasis","family":"Chakraborty","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6180-350X","authenticated-orcid":false,"given":"Daniel","family":"Galea","sequence":"additional","affiliation":[]}],"member":"266","published-online":{"date-parts":[[2025,12,18]]},"reference":[{"key":"mlstae281ebib1","doi-asserted-by":"publisher","first-page":"108","DOI":"10.1016\/j.rse.2012.04.024","type":"journal-article","article-title":"Estimating air surface temperature in Portugal using MODIS LST data","volume":"124","author":"Benali","year":"2012","journal-title":"Remote Sens. Environ."},{"key":"mlstae281ebib2","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1561\/2200000006","type":"journal-article","article-title":"Learning deep architectures for AI","volume":"2","author":"Bengio","year":"2009","journal-title":"Found. Trends\u00ae Mach. Learn."},{"key":"mlstae281ebib3","doi-asserted-by":"publisher","first-page":"2391","DOI":"10.1175\/JHM-D-19-0232.1","type":"journal-article","article-title":"A comparison of the US climate reference network precipitation data to the Parameter-Elevation Regressions on Independent Slopes Model (PRISM)","volume":"21","author":"Buban","year":"2020","journal-title":"J. Hydrometeorol."},{"key":"mlstae281ebib4","first-page":"17164","type":"conference-proceedings","article-title":"Calibrated reliable regression using maximum mean discrepancy","volume":"vol 33","author":"Cui","year":"2020"},{"key":"mlstae281ebib5","doi-asserted-by":"publisher","first-page":"2031","DOI":"10.1002\/joc.1688","type":"journal-article","article-title":"Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States","volume":"28","author":"Daly","year":"2008","journal-title":"Int. J. Climatol."},{"key":"mlstae281ebib6","doi-asserted-by":"publisher","first-page":"197","DOI":"10.1561\/2000000039","type":"journal-article","article-title":"Deep learning: methods and applications","volume":"7","author":"Deng","year":"2014","journal-title":"Found Trends\u00ae Signal Process."},{"key":"mlstae281ebib7","doi-asserted-by":"publisher","first-page":"527","DOI":"10.1007\/s00382-010-0977-x","type":"journal-article","article-title":"Uncertainty in climate change projections: the role of internal variability","volume":"38","author":"Deser","year":"2012","journal-title":"Clim. Dyn."},{"key":"mlstae281ebib8","doi-asserted-by":"publisher","first-page":"963","DOI":"10.1038\/s41561-024-01527-w","type":"journal-article","article-title":"AI-empowered next-generation multiscale climate modelling for mitigation and adaptation","volume":"17","author":"Eyring","year":"2024","journal-title":"Nat. Geosci."},{"key":"mlstae281ebib9","doi-asserted-by":"publisher","DOI":"10.1029\/2007JD008470","type":"journal-article","article-title":"A global monthly land surface air temperature analysis for 1948\u2013present","volume":"113","author":"Fan","year":"2008","journal-title":"J. Geophys. Res."},{"key":"mlstae281ebib10","doi-asserted-by":"publisher","DOI":"10.1016\/j.atmosres.2025.108015","type":"journal-article","article-title":"Impact of snow cover on the surface energy balance and its contribution to the extreme cold wave in Spain after snowstorm Filomena","volume":"320","author":"Fern\u00e1ndez-Castillo","year":"2025","journal-title":"Atmos. Res."},{"key":"mlstae281ebib11","doi-asserted-by":"publisher","first-page":"1033","DOI":"10.1175\/2008JAMC2084.1","type":"journal-article","article-title":"Downscaling climate over complex terrain: high finescale (<1000 m) spatial variation of near-ground temperatures in a montane forested landscape (Great Smoky Mountains)","volume":"48","author":"Fridley","year":"2009","journal-title":"J. Appl. Meteorol. Climatol."},{"key":"mlstae281ebib12","first-page":"1050","type":"conference-proceedings","article-title":"Dropout as a Bayesian approximation: representing model uncertainty in deep learning","author":"Gal","year":"2016"},{"key":"mlstae281ebib13","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1111\/j.1467-9868.2007.00587.x","type":"journal-article","article-title":"Probabilistic forecasts, calibration and sharpness","volume":"69","author":"Gneiting","year":"2007","journal-title":"J. R. Stat. Soc. Ser. B Stat. Methodol."},{"key":"mlstae281ebib14","doi-asserted-by":"crossref","first-page":"568","DOI":"10.1038\/s41586-019-1559-7","type":"journal-article","article-title":"Deep learning for multi-year ENSO forecasts","volume":"573","author":"Ham","year":"2019","journal-title":"Nature"},{"key":"mlstae281ebib15","first-page":"1026","type":"conference-proceedings","article-title":"Delving deep into rectifiers: surpassing human-level performance on imagenet classification","author":"He","year":"2015"},{"key":"mlstae281ebib16","doi-asserted-by":"publisher","first-page":"1999","DOI":"10.1002\/qj.3803","type":"journal-article","article-title":"The ERA5 global reanalysis","volume":"146","author":"Hersbach","year":"2020","journal-title":"Q. J. R. Meteorol. Soc."},{"key":"mlstae281ebib17","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.102.103509","type":"journal-article","article-title":"Parameter estimation for the cosmic microwave background with Bayesian neural networks","volume":"102","author":"Hort\u00faa","year":"2020","journal-title":"Phys. Rev. D"},{"article-title":"Bayesian-Torch: Bayesian neural network layers for uncertainty estimation","year":"2020","author":"Krishnan","key":"mlstae281ebib18","type":"other"},{"key":"mlstae281ebib19","first-page":"344","type":"book","article-title":"Learnable uncertainty under Laplace approximations","author":"Kristiadi","year":"2021"},{"key":"mlstae281ebib20","first-page":"11683","type":"conference-proceedings","article-title":"Calibrated and sharp uncertainties in deep learning via density estimation","author":"Kuleshov","year":"2022"},{"key":"mlstae281ebib21","doi-asserted-by":"publisher","first-page":"1459","DOI":"10.1007\/s12145-023-00970-4","type":"journal-article","article-title":"On the modern deep learning approaches for precipitation downscaling","volume":"16","author":"Kumar","year":"2023","journal-title":"Earth Sci. Inf."},{"key":"mlstae281ebib22","type":"conference-proceedings","article-title":"Simple and scalable predictive uncertainty estimation using deep ensembles","volume":"vol 30","author":"Lakshminarayanan","year":"2017"},{"key":"mlstae281ebib23","type":"book","article-title":"Convolutional networks for images, speech, and time series","volume":"vol 3361","author":"LeCun","year":"1995"},{"key":"mlstae281ebib24","first-page":"253","type":"conference-proceedings","article-title":"Convolutional networks and applications in vision","author":"LeCun","year":"2010"},{"key":"mlstae281ebib25","doi-asserted-by":"publisher","DOI":"10.1175\/AIES-D-24-0095.1","type":"journal-article","article-title":"Evaluating probabilistic deep learning methods for uncertainty quantification of precipitation bias correction","volume":"4","author":"Lops","year":"2025","journal-title":"Artif. Intell. Earth Syst."},{"key":"mlstae281ebib26","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s13143-023-00323-7","type":"journal-article","article-title":"Development of deep convolutional neural network ensemble models for 36-month ENSO forecasts","author":"Lops","year":"2023","journal-title":"Asia-Pac. J. Atmos. Sci."},{"key":"mlstae281ebib27","first-page":"457","type":"conference-proceedings","article-title":"Transformer for single image super-resolution","author":"Lu","year":"2022"},{"key":"mlstae281ebib28","doi-asserted-by":"publisher","first-page":"124","DOI":"10.1038\/s43247-025-02042-5","type":"journal-article","article-title":"Residual corrective diffusion modeling for km-scale atmospheric downscaling","volume":"6","author":"Mardani","year":"2025","journal-title":"Commun. Earth Environ."},{"key":"mlstae281ebib29","doi-asserted-by":"publisher","DOI":"10.1029\/2010JG001486","type":"journal-article","article-title":"A global comparison between station air temperatures and MODIS land surface temperatures reveals the cooling role of forests","volume":"116","author":"Mildrexler","year":"2011","journal-title":"J. Geophys. Res."},{"key":"mlstae281ebib30","doi-asserted-by":"publisher","first-page":"1419","DOI":"10.1002\/2013JD019880","type":"journal-article","article-title":"Orographic effects on snow deposition patterns in mountainous terrain","volume":"119","author":"Mott","year":"2014","journal-title":"J. Geophys. Res."},{"key":"mlstae281ebib31","doi-asserted-by":"publisher","first-page":"7324","DOI":"10.1073\/pnas.0808533106","type":"journal-article","article-title":"Spatial and seasonal patterns in climate change, temperatures, and precipitation across the United States","volume":"106","author":"Portmann","year":"2009","journal-title":"Proc. Natl Acad. Sci."},{"key":"mlstae281ebib32","doi-asserted-by":"crossref","first-page":"85","DOI":"10.1016\/j.neunet.2014.09.003","type":"journal-article","article-title":"Deep learning in neural networks: an overview","volume":"61","author":"Schmidhuber","year":"2015","journal-title":"Neural Netw."},{"key":"mlstae281ebib33","doi-asserted-by":"publisher","first-page":"270","DOI":"10.1038\/s41612-025-01157-y","type":"journal-article","article-title":"A generative framework for probabilistic, spatiotemporally coherent downscaling of climate simulation","volume":"8","author":"Schmidt","year":"2025","journal-title":"npj Clim. Atmos. Sci."},{"key":"mlstae281ebib34","first-page":"1874","type":"conference-proceedings","article-title":"Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network","author":"Shi","year":"2016"},{"key":"mlstae281ebib35","doi-asserted-by":"publisher","first-page":"1215","DOI":"10.1175\/JHM-D-18-0041.1","type":"journal-article","article-title":"Downscaling of ERA-interim temperature in the contiguous United States and its implications for rain\u2013snow partitioning","volume":"19","author":"Tang","year":"2018","journal-title":"J. Hydrometeorol."},{"key":"mlstae281ebib36","doi-asserted-by":"publisher","first-page":"3953","DOI":"10.5194\/gmd-16-3953-2023","type":"journal-article","article-title":"The fully coupled regionally refined model of E3SM version 2: overview of the atmosphere, land, and river results","volume":"16","author":"Tang","year":"2023","journal-title":"Geosci. Model Dev."},{"key":"mlstae281ebib37","doi-asserted-by":"publisher","first-page":"4639","DOI":"10.5194\/amt-14-4639-2021","type":"journal-article","article-title":"Assessing sub-grid variability within satellite pixels over urban regions using airborne mapping spectrometer measurements","volume":"14","author":"Tang","year":"2021","journal-title":"Atmos. Meas. Tech."},{"key":"mlstae281ebib38","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ab7e1a","type":"journal-article","article-title":"Methods for comparing uncertainty quantifications for material property predictions","volume":"1","author":"Tran","year":"2020","journal-title":"Mach. Learn."},{"key":"mlstae281ebib39","doi-asserted-by":"publisher","DOI":"10.1029\/2005GL022760","type":"journal-article","article-title":"Relationships between precipitation and surface temperature","volume":"32","author":"Trenberth","year":"2005","journal-title":"Geophys. Res. Lett."},{"key":"mlstae281ebib40","first-page":"1663","type":"conference-proceedings","article-title":"Deepsd: generating high resolution climate change projections through single image super-resolution","author":"Vandal","year":"2017"},{"key":"mlstae281ebib41","doi-asserted-by":"publisher","first-page":"1417","DOI":"10.1175\/JCLI-D-16-0168.1","type":"journal-article","article-title":"Incorporating snow albedo feedback into downscaled temperature and snow cover projections for California\u2019s Sierra Nevada","volume":"30","author":"Walton","year":"2017","journal-title":"J. Clim."},{"article-title":"Flipout: efficient pseudo-independent weight perturbations on mini-batches","year":"2018","author":"Wen","key":"mlstae281ebib42","type":"preprint"},{"key":"mlstae281ebib43","doi-asserted-by":"publisher","DOI":"10.1175\/AIES-D-23-0076.1","type":"journal-article","article-title":"UFNet: joint U-Net and fully connected neural network to bias correct precipitation predictions from climate models","volume":"3","author":"Yu","year":"2024","journal-title":"Artif. Intell. Earth Syst."}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,18]],"date-time":"2025-12-18T09:09:59Z","timestamp":1766048999000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ae281e"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,18]]},"references-count":43,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2025,12,18]]},"published-print":{"date-parts":[[2025,12,30]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/ae281e","relation":{},"ISSN":["2632-2153"],"issn-type":[{"type":"electronic","value":"2632-2153"}],"subject":[],"published":{"date-parts":[[2025,12,18]]},"assertion":[{"value":"Evaluating probabilistic deep learning methods for uncertainty quantification of temperature downscaling","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2025 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2025-04-18","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-12-04","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-12-18","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}