SOSeas Web App: An assessment web-based decision support tool to predict dynamic risk of drowning on beaches using deep neural networks

People still drown on beaches in unacceptable numbers due to the lack of knowledge about the risks taking place in them. The proposed methodology forecasts electronic bathing flags in beaches by integrating the benefits of metocean operational systems, machine learning and web-based decision support...

Full description

Saved in:
Bibliographic Details
Published in:Journal of operational oceanography 2023-05, Vol.16 (2), p.155-174
Main Authors: García-Alba, Javier, Bárcena, Javier F., Pedraz, Luis, Fernández, Felipe, García, Andrés, Mecías, Marcos, Costas-Veigas, Javier, Samano, María Luisa, Szpilman, David
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:People still drown on beaches in unacceptable numbers due to the lack of knowledge about the risks taking place in them. The proposed methodology forecasts electronic bathing flags in beaches by integrating the benefits of metocean operational systems, machine learning and web-based decision support technologies into a 24/7 risk assessment service that could be easily implemented at any beach worldwide with low costs of maintenance. Firstly, a crosscutting analysis between metocean conditions, beach characteristics and flag records was performed. Secondly, an expert system, based on Deep Learning, was developed to obtain electronic bathing flags as an indicator of the dynamic risk of drowning on beaches. The input variables of the Deep Neural Network were significant wave height, mean wave period, wind velocity, marine current velocity, incidence angle, and beach modal state. Finally, the application of the method to the Santa Catarina's beaches (Brazil) conveniently reproduced the status flag of beaches.
ISSN:1755-876X
1755-8778
DOI:10.1080/1755876X.2021.1999107