CRBM-DBN-based prediction effects inter-comparison for significant wave height with different patterns. (15th September 2021)

• Record Type: Journal Article
• Authors: Dai, Hao; Shang, Shaoping; Lei, Famei; Liu, Ke; Zhang, Xining; Wei, Guomei; Xie, Yanshuang; Yang, Shuai; Lin, Rui; Zhang, Weijie
• Is Part Of: Ocean engineering. Volume 236(2021)
• Publication Date: 2021-09-15
• Subjects: CRBM-DBN -- Gulf of Mexico -- Prediction for the significant wave height -- Univariate and multivariate prediction patterns; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2021.109559
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Based on the Conditional Restricted Boltzmann Machine - Deep Belief Network (CRBM-DBN), we select four patterns and compare their prediction effects for the significant wave height in the Gulf of Mexico (GoM). Historical datasets of all 12 buoys managed by the National Data Buoy Center are employed to train and construct models. Root-mean-square error (RMSE) and coefficient of efficiency (CE) between the observed and predicted wave heights are investigated. We find that for the short-term prediction (i.e., lead time≤12 h), the best results (RMSE<0.24 m and CE > 0.92) are achieved with the univariate significant wave height as the input in most cases of the whole gulf. When the lead time is equal to 24 h or 48 h, the multivariate pattern of "significant wave height + dominant wave direction + wind speed + wind direction" has the optimal effects (0.18 m < RMSE<0.40 m and 0.72 < CE < 0.93) in the vicinity of 26 。 N oceans. The superiority is very obvious and gradually diminishes as the latitude increases to the north and decreases to the south. Regarding the wave height predictions in different oceans of GoM, the findings provide evidence that it may be contributed to select optimal prediction patterns and obtain the best applications. Highlights: Based on the CRBM-DBN, we investigate the prediction effects of four patterns. Univariant wave height, and multivariant wave and wind patterns have respective optimal effects. The research results are helpful to select the best patterns for different predictive requirements.