Logotipo do repositório

Machine learning-based hydrological models for flash floods: a systematic literature review

dc.contributor.authorSantos, Leonardo B. L.
dc.contributor.authorEscobar-Silva, Elton V.
dc.contributor.authorSatolo, Luiz F.
dc.contributor.authorOyarzabal, Ricardo S.
dc.contributor.authorDiniz, Michael M.
dc.contributor.authorNegri, Rogério G. [UNESP]
dc.contributor.authorLima, Glauston R. T.
dc.contributor.authorStephany, Stephan
dc.contributor.authorSoares, Jaqueline A. J. P.
dc.contributor.authorDuque, Johan S.
dc.contributor.authorSaraiva Filho, Fernando L.
dc.contributor.authorCaseri, Angélica N.
dc.contributor.authorBacelar, Luiz
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)pt
dc.date.accessioned2026-08-28T18:33:36Z
dc.date.issued2025-10-11
dc.description.abstractFlash floods are critical events for emergency management, yet their modeling remains highly challenging, even in smart cities approaches. Physically based hydrological models are often unsuitable at small spatiotemporal scales due to their computational complexity and dependence on detailed local parameters, which are rarely available during flash floods. With the growing availability of hydrological data, machine learning (ML) has emerged as a promising alternative. This work performs a Systematic Literature Review (SLR) to improve our understanding of the research landscape on ML applications for flash flood forecasting, a significant subset of flash flood modeling. From more than 1,200 papers published until January 2024 in Web of Science, SCOPUS/Elsevier, Springer/Nature, and Wiley, 50 were selected following PRISMA guidelines. The inclusion and exclusion criteria removed reviews, retractions, papers focused on post-flood damage assessment (not forecasting), and those with time resolutions of 6 hours or more, retaining only studies with fine-scale temporal data (<6 hours). For each paper, we extracted information on forecasting horizon, study area size, input data, ML techniques, and outcomes (regression or classification). Results show a sharp rise in ML-based flash flood research, with China leading (38%). Nearly all studies rely on rainfall, discharge, and water level data - often in combination. Long short-term memory (LSTM) networks dominate (60%). Unfortunately, only 10% of the selected studies provide access to their datasets. This lack of transparency poses a major barrier to reproducibility, inhibits fair comparative evaluation of models, and ultimately slows methodological progress in flash flood forecasting. Furthermore, our review highlights that no method consistently outperforms others. This variability in performance is likely influenced by factors such as regional hydrological characteristics (e.g., differences between arid and tropical basins), variations in input data quality, and the length of the forecast horizon (e.g., 1- vs. 6-hour prediction). Lastly, we recommend advancing this field through integration with early warning systems, creation of benchmarks, open data practices, and stronger multidisciplinary collaboration.
dc.description.affiliationNational Center for Monitoring and Early Warning of Natural Disasters, Cemaden, Estrada Dr. Altino Bondensan, 500, 12247-016, São José dos Campos, São Paulo, Brazil
dc.description.affiliationNational Institute for Space Research, INPE, Av. dos Astronautas, 1758, 12227-010, São José dos Campos, São Paulo, Brazil
dc.description.affiliationFederal University of São Paulo, UNIFESP, Avenida Cesare Mansueto Giulio Lattes, 1201, 12247-014, São José dos Campos, São Paulo, Brazil
dc.description.affiliationFederal Institute of São Paulo, IFSP, Rod. Presidente Dutra, km 145, 12223-201, São José dos Campos, São Paulo, Brazil
dc.description.affiliationSão Paulo State University, UNESP, Rod. Presidente Dutra, km 137.8, 12247-004, São José dos Campos, São Paulo, Brazil
dc.description.affiliationTechnological University of Uruguay, UTEC, Maciel sn/esq Morquio, Durazno, 97000, Durazno, Durazno Department, Uruguay
dc.description.affiliationUniversity of São Paulo, Av. Trabalhador São-carlense, 400, 13566-590, São Carlos, São Paulo, Brazil
dc.description.affiliationDuke University, Science Dr., 27710, Durham, North Carolina, USA
dc.description.affiliationUnespSão Paulo State University, UNESP, Rod. Presidente Dutra, km 137.8, 12247-004, São José dos Campos, São Paulo, Brazil
dc.identifierhttps://app.dimensions.ai/details/publication/pub.1193823664
dc.identifier.dimensionspub.1193823664
dc.identifier.doi10.1007/s44268-025-00071-9
dc.identifier.issn2731-9032
dc.identifier.orcid0000-0002-9437-9351
dc.identifier.orcid0000-0001-5586-6424
dc.identifier.orcid0000-0003-3024-2062
dc.identifier.orcid0000-0003-3853-8749
dc.identifier.orcid0000-0002-4808-2362
dc.identifier.orcid0000-0002-6302-4259
dc.identifier.orcid0000-0002-7940-8804
dc.identifier.orcid0000-0002-0703-3986
dc.identifier.orcid0000-0001-8389-8028
dc.identifier.urihttps://hdl.handle.net/11449/331184
dc.publisherSpringer Nature
dc.relation.ispartofSmart Construction and Sustainable Cities; n. 1; v. 3; p. 21
dc.rights.accessRightsAcesso abertopt
dc.rights.sourceRightsoa_all
dc.rights.sourceRightsgold
dc.sourceDimensions
dc.titleMachine learning-based hydrological models for flash floods: a systematic literature review
dc.typeArtigopt
dspace.entity.typePublication
relation.isOrgUnitOfPublicationc73b286a-b5fa-4312-a7ec-62f987e7b514
relation.isOrgUnitOfPublication.latestForDiscoveryc73b286a-b5fa-4312-a7ec-62f987e7b514
unesp.campusUniversidade Estadual Paulista (UNESP), Instituto de Ciência e Tecnologia, São José dos Campospt

Arquivos

Pacote original

Agora exibindo 1 - 1 de 1
Carregando...
Imagem de Miniatura
Nome:
fulltext_11449_331184.pdf
Tamanho:
1,68 MB
Formato:
Unknown data format
Descrição:
Obtido de: Open Alex