Calibration and skill assessment of two input and dissipation parameterizations in WAVEWATCH-III model forced with ERA5 winds with application to Persian Gulf and Gulf of Oman

Abstract

The wind energy input and whitecap dissipation terms in the third generation wave models have been improved over the time. In this research, the well-known WAM-Cycle4, known as ST3, and the newest ST6 package in WAVEWATCH-III model were employed combined with ERA5 wind data over the Persian Gulf and Gulf of Oman. In order to model the wave height appropriately over the study area in comparison with both in-situ and satellite observations, new combined error parameters were introduced leading to two major calibration values: (1) the local calibration values relating to either Persian Gulf (PGc) or Gulf of Oman (GOc); (2) the global calibration values over the entire computational domain (PGGOc). Considerably different model performance was appeared in the Persian Gulf (PG) compared to either the Gulf of Oman (GO) or entire computational domain (PGGO) when ST3 package was employed, while obtained results using ST6 package showed negligible sensitivity to the domain. Both ST3 and ST6 packages with default tuning values significantly underestimated the wave height and peak period in PG; however, slight underestimation (~0.03 m) for wave height and overestimation (~0.1 s) for mean wave period were obtained in GO. During the time period the cyclone Phet has traversed across the study area, the model overestimated the wave height in comparison with altimeter dataset. Mean bias distribution over the study area indicates that calibrated ST6 outperforms calibrated ST3 over GO; however, both packages provide similar results over PG. Considering different methods for nonlinear interaction term led to similar directional wave spectrums over the study area and negligible changes in bulk wave parameters.