A Probabilistic Approach to Estimate Design Wave Parameters and Extreme Wave Return Values for 100 Years in the Indian Ocean

Abstract

Unexpected, huge, and dangerous rogue or freak waves on calm sea surfaces still remain unpredicted causing destruction to ships and human lives. Probabilistic wave modeling is an approach for long-term operational and extreme wave parameter estimation. Using extreme value distributions like Generalized Extreme Value Distribution (GEV), Generalized Pareto Distribution (GPD), and Weibull Distribution (WD), mathematical equations have been estimated, for the extreme wave height for given return periods, mean maximum wave height, and most frequent wave heights. Using NCEP daily wind data, 37 years (1981–2017) model computed wave height data (SWH) is considered for a grid (72° E, 9° N) in the Indian Ocean through which OCKHI cyclone passed in 2017. The model computed and the distribution of the estimated extreme wave heights were compared over some given return periods 5, 10, 25, 50, and 100 years. The results are not satisfactory due to the coarse and smoothed SWH time series data. To improve the results for the same above grid, NOAA WAVEWATCH III model generated SWH data from 2006–2017 are considered next having better accuracy and finer resolution. At the above location, for a return period of 5 years, the model computed SWH value is 5.56 m, and the GPD, WD, and GEV estimated values are 5.42, 5.59, and 1.89 m, respectively. On comparing computed and estimated extreme wave heights for both normal and cyclonic periods for given return periods, it was found that GPD performed best followed by WD. Different distributions performed with varied accuracy while estimating the mean maximum wave height and most frequent wave heights.