[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124452-en":3,"doc-seo-124452-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},124452,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",8,"Research & Report","Coupling machine learning and physical modelling for predicting runoff at catchment scale","This paper proposes a hybrid approach combining data-driven learning with physical modelling to predict runoff occurrence and runoff volume at catchment scale. Runoff volume is first estimated from recorded storms using the Green-Ampt infiltration model. LightGBM and deep neural network models are then trained to predict outputs of the physical model using atmospheric variables collected before or immediately after storm onset. Results from a small urban catchment in Madrid show DNN achieves better prediction accuracy, and adding auxiliary storm-intensity and rain estimates substantially improves performance. The study presents physics-informed data-driven algorithms that learn from physical-model outputs, moving beyond common hydrological ML practices.","Journal of Environmental Management 354 (2024) 120404  \nContents lists available at ScienceDirect  \nJournal of Environmental Management  \njournal [homepage: www.elsevier.com/locate/jenvman](homepage: www.elsevier.com/locate/jenvman)  \n| Research article\u003Cbr>Coupling machine learning and physical modelling for predicting runoff at catchment scale |  |  |  |\n| --- | --- | --- | --- |\n| Sergio Zubelzua, *, Abdulmomen Ghalkhab, Chaouki Ben Issaid b, Andrea Zanellac, Medhi Bennisb\u003Cbr>a Departamento de Ingeniería Agroforestal, Universidad Polit´ecnica de Madrid, Madrid, Spain b Faculty of Information Technology and Electrical Engineering, University of Oulu, Oulu, Finland c Department of Information Engineering, University of Padova, Padova, Italy |  |  |  |\n| A R T I C L E I N F O |  | A B S T R A C T |  |\n| Handling Editor: Jason Michael Evans |  | In this paper, we present an approach that combines data-driven and physical modelling for predicting the runoff occurrence and volume at catchment scale. With that aim, we first estimated the runoff volume from recorded storms aided by the Green-Ampt infiltration model. Then, we used machine learning algorithms, namely LightGBM (LGBM) and Deep Neural Network (DNN), to predict the outputs of the physical model fed on a set of atmospheric variables (relative humidity, temperature, atmospheric pressure, and wind velocity) collected before or immediately after the beginning of the storm. Results for a small urban catchment in Madrid show DNN performed better in predicting the runoff occurrence and volume. Moreover, enriching the input primary atmospheric variables with auxiliary variables (e.g., storm intensity data recorded during the first hour, or rain volume and intensity estimates obtained from auxiliary regression methods) largely increased the model performance. We show in this manuscript data-driven algorithms shaped by physical criteria can be successfully generated by allowing the data-driven algorithm learn from the output of physical models. It represents a novel approach for physics-informed data-driven algorithms shifting from common practices in hydrological modelling through machine learning. |  |\n| Keywords: Hydrology Machine learning\u003Cbr>Physical modelling Runoff Catchment |  |  |  |\n\n1. Introduction  \nIn 2017, a set of 230 experts, including well known hydrologists and scientists from other related disciplines, highlighted several unsolved questions in hydrology (Bl¨osch et al., 2017). They arose some issues related to hydrologic laws and their suitability at different scales, the use of historical data vs soft data, the reduction of the amount of model structural/parameter/input uncertainty in hydrological prediction, among others. Following the conclusions presented in that work, the comprehensive understanding of the complex interactions among the physical processes underlying the hydrological systems still remained elusive.  \nThe accurate modelling of hydrological systems is complex, involving highly variable and interlinked distant processes, such as precipitation, infiltration and flood routing. Theories for modelling such phenomena were proposed long time ago (see, for example, Thornthwaite and Holzman, 1939; Darcy, 1856; Philip, 1957; Richards,  \n1931; Saint Venant, 1871). Nonetheless, physical models still suffer from major problems, such as mathematical complexity, unreliability of parameterization when modelling large and heterogeneous watersheds, inability to adequately handle extreme spatial and temporal variability, or lack of random criteria in modelling the evolution of hydrological systems. Physical models are hence rigorous from a conceptual perspective but often prohibitively complex when large-scale accurate results are sought.  \nGiven the performance achieved by data-driven models in different fields, and the always larger amount of data available, many hydrologists have then explored in the last years the suitability of applying machine-learning (ML) ","cbCaisqK3c9asq0D","https://ap.wps.com/l/cbCaisqK3c9asq0D","pdf",3678538,1,9,"English","en",105,"# Introduction\n## Background and motivations in hydrology\n## Limits of physical models and opportunities for ML\n## Two ML pathways: autoregressive vs physics-informed approaches","[{\"question\":\"How does the proposed method combine physical modelling with machine learning?\",\"answer\":\"It first estimates runoff volume with the Green-Ampt infiltration model, then trains machine learning models to predict outputs of that physical model using selected atmospheric variables collected before or right after storm onset.\"},{\"question\":\"Which machine learning models are used, and what performs best?\",\"answer\":\"The approach uses LightGBM and a deep neural network. For a small urban catchment in Madrid, the deep neural network performs better for predicting runoff occurrence and volume.\"},{\"question\":\"Why does enriching input variables improve model performance?\",\"answer\":\"Adding auxiliary variables—such as storm intensity in the first hour and rain volume/intensity estimates from auxiliary regression methods—largely increases overall prediction performance.\"}]","Coupling machine learning and physical modelling for predicting runoff at catchment scale | PDF",1785822364,23,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"coupling-machine-learning-and-physical-modelling-for-predicting-runoff-at-catchment-scale","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/coupling-machine-learning-and-physical-modelling-for-predicting-runoff-at-catchment-scale/124452/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"How does the proposed method combine physical modelling with machine learning?","Question",{"text":75,"@type":76},"It first estimates runoff volume with the Green-Ampt infiltration model, then trains machine learning models to predict outputs of that physical model using selected atmospheric variables collected before or right after storm onset.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which machine learning models are used, and what performs best?",{"text":80,"@type":76},"The approach uses LightGBM and a deep neural network. For a small urban catchment in Madrid, the deep neural network performs better for predicting runoff occurrence and volume.",{"name":82,"@type":73,"acceptedAnswer":83},"Why does enriching input variables improve model performance?",{"text":84,"@type":76},"Adding auxiliary variables—such as storm intensity in the first hour and rain volume/intensity estimates from auxiliary regression methods—largely increases overall prediction performance.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]