Analysis of Meteorological Multi-Factor Influences on Wave Height Based on Machine Learning
LIN Xiaofeng
NIE Huiwen
Abstract：To understand the accuracy of machine learning in wave prediction,this study utilizes hourly observational data from the Maoming buoy station in Guangdong from February 2013 to June 2024.The SHAP(SHapley Additive exPlanations)machine learning method is applied to analyze the specific influ-ences of meteorological factors such as wind speed,air temperature,and air pressure on significant wave height(SWH)in the coastal waters of Guangdong.Additionally,the performance characteristics of ten ma-chine learning models in predicting SWH are evaluated.The results indicate the following.Wind speed is the most critical factor in SWH prediction,followed by air temperature and month.In two experimental sce-narios-single-factor(wind speed)and multi-factor(wind speed,month,air temperature)model fitting-support vector regression(SVR)consistently outperforms other models.Particularly in multi-factor fitting,SVR achieves a root mean square error(RMSE)of approximately 0.31,a correlation coefficient(CORR)of 0.87,and a coefficient of determination(R2)of 0.74,demonstrating excellent predictive performance.Light gradient boosting machine regression(LGBM)exhibits comparable stability to SVR in prediction.
Keywords：marine meteorologysignificant wave heightmachine learningmeteorological factorssupport vector regressionlight gradient boosting machine regression
Publication Date：2025-08-20
Online Publishing Date：2025-09-15(First online date of this platform, not the publication date of the document)
Pages：5( 33-37 )