Learning optimal CNN architectures for Precipitation Downscaling with Differentiable Architecture Search

Learning optimal CNN architectures for Precipitation Downscaling with Differentiable Architecture Search | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 21 August 2025 V1 Latest version Share on Learning optimal CNN architectures for Precipitation Downscaling with Differentiable Architecture Search Authors : Leonardo Alessandro Lüder 0009-0008-8651-813X , Mikhail Ivanov , and Ramon Fuentes Franco 0000-0002-3085-0175 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.175580225.56700263/v1 201 views 194 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Climate change poses a significant threat to both ecosystems and human societies, partly due to the projected increase in extreme precipitation events (EPEs). Accurately identifying where such events are likely to occur more frequently is essential for minimizing societal and economic impacts. Although global climate models (GCMs) provide reliable long-term projections, their coarse spatial resolution limits their ability to resolve localized extremes. Convolutional neural networks (CNNs) with upsampling layers can be used to statistically downscale GCM outputs to finer spatial resolutions. However, their architectures are typically designed manually. This paper introduces a modular and automated framework based on Differentiable Architecture Search (DARTS) to discover high-performing CNNs for precipitation downscaling. The search space is intentionally designed to remain close to that of the baseline model, ensuring fair comparison while allowing systematic exploration of different cell and node configurations. Results show that all sufficiently complex DARTS-generated architectures outperform the baseline in terms of validation loss, Matthews correlation coefficient, and diagnostic odds ratio. The discovered models also generalize well to a future climate scenario (SSP370), indicating robustness beyond the training distribution. Overall, this study demonstrates that automated and modular architecture search can reliably produce effective CNNs for statistical downscaling. The proposed framework reduces reliance on manual architecture design and lowers the barrier for deploying deep learning in climate services. Supplementary Material File (1044364_0_merged_1755010857.pdf) Download 2.14 MB File (article_file.pdf) Download 1.13 MB File (tables.pdf) Download 63.97 KB Information & Authors Information Version history V1 Version 1 21 August 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords climatology (global change) darts downscaling extreme precipitation neural networks optimization precipitation Authors Affiliations Leonardo Alessandro Lüder 0009-0008-8651-813X Sveriges meteorologiska och hydrologiska institut View all articles by this author Mikhail Ivanov Sveriges meteorologiska och hydrologiska institut View all articles by this author Ramon Fuentes Franco 0000-0002-3085-0175 [email protected] Sveriges meteorologiska och hydrologiska institut View all articles by this author Funding Information European Commission 10114295, 10103109, 10093450 Metrics & Citations Metrics Article Usage 201 views 194 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Leonardo Alessandro Lüder, Mikhail Ivanov, Ramon Fuentes Franco. Learning optimal CNN architectures for Precipitation Downscaling with Differentiable Architecture Search. Authorea . 21 August 2025. DOI: https://doi.org/10.22541/au.175580225.56700263/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.175580225.56700263/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fef699a5bfd4193',t:'MTc3OTMyMjc5Nw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();