Using machine learning to derive spatial wave data: A case study for a marine energy site. (August 2021)

• Record Type: Journal Article
• Authors: Chen, Jiaxin; Pillai, Ajit C.; Johanning, Lars; Ashton, Ian
• Is Part Of: Environmental modelling & software. Volume 142(2021)
• Publication Date: 2021-08
• Subjects: Nearshore wave modelling -- Random forest -- Machine learning -- Spatial prediction -- Optimal gridding; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2021.105066
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: Ocean waves are widely estimated using physics-based computational models, which predict how energy is transferred from the wind, dissipated, and transferred spatially across the ocean. Machine learning methods offer an opportunity to predict these data with significantly reduced data input and computational power. This paper describes a novel surrogate model developed using the random forest method, which replicates the spatial nearshore wave data estimated by a Simulating WAves Nearshore (SWAN) numerical model. By incorporating in-situ buoy observations, outputs were found to match observations at a test location more closely than the corresponding SWAN model. Furthermore, the required computational time reduced by a factor of 100. This methodology can provide accurate spatial wave data in situations where computational power and transmission are limited, such as autonomous marine vehicles or during coastal and offshore operations in remote areas. This represents a significant supplementary service to existing physics-based wave models. Highlights: A machine learning method for constructing a surrogate SWAN wave model is presented. Model demonstrates machine learning of spatial correlation of ocean waves. Optimal spatial gridding to represent high resolution wave distribution. Using observations as input provides accurate real-time wave model output. Surrogate model outperforms SWAN model during validation.