Error distribution and correction of the predicted wave characteristics over the Persian Gulf

• Published in: Ocean engineering 2014-01, Vol.75, p.81-89
• Main Authors: Moeini, Mohammad Hadi, Etemad-Shahidi, Amir, Chegini, Vahid, Rahmani, Iraj, Moghaddam, Mona
• Format: Article
• Language: English
• Subjects: Earth, ocean, space; Error distribution; Exact sciences and technology; External geophysics; Persian Gulf; Physics of the oceans; Sea-air exchange processes; SWAN; Wave hindcasting; Wave modeling
• ISSN: 0029-8018; 1873-5258
• DOI: 10.1016/j.oceaneng.2013.11.012

Wind-waves are the most important environmental parameter for the design of coastal and offshore structures, sediment transport, coastal erosion etc. Therefore, an accurate evaluation of the wave climate is of great importance. Due to the lack of long term measurements, nowadays numerically modeled wave data are widely used for determining the wave climate. The numerically simulated wave data are continuous in time and space, but generally inaccurate in enclosed and semi-enclosed basins mainly due to the inaccurate wind input data. The main goal of this study is to develop a new and efficient approach to improve the hindcasted wave parameters in the Persian Gulf. Hence, the third generation SWAN model was employed for the wave modeling forced by the 6-hourly ECMWF wind data with a resolution of 0.5°. A new methodology was introduced for the distribution of wave prediction errors from discrete observation points to the other points of interest. It was found that the proposed method which considers the wave generation physics, leads to a significant improvement in the predicted wave parameters. In addition, it was revealed that the improvements in prediction of waves with higher wave heights and longer periods are more than those of others. This was shown to be due to the higher correlation between high values of output parameters which contain larger errors. The influence radius in the error distribution procedure was found to be near 2° (~200km). •The SWAN model was used for the wave modeling forced by the 6-hourly ECMWF wind data.•A new methodology was introduced for the distribution of wave prediction errors.•Two approaches for error distribution and improvement of the results were assessed.•The influence factor in the updating procedure was found to be near 2°.