CFE 2020: Start Registration
View Submission - CMStatistics
B0549
Title: Time series parameter estimation for ocean wave models Authors:  Jake Grainger - Lancaster University (United Kingdom) [presenting]
Adam Sykulski - Lancaster University (United Kingdom)
Philip Jonathan - Lancaster University / Shell Research Limited (United Kingdom)
Kevin Ewans - Met Ocean Research Limited (New Zealand)
Abstract: Understanding the behaviour of wind-generated ocean waves is important to many offshore and coastal engineering activities. Estimating models for the frequency domain behaviour of wind-generated wave time series has received considerable attention in the oceanographic literature. Typically, parametric spectral forms (such as JONSWAP) are fitted to periodograms calculated from observed time series, using least-squares techniques. We demonstrate, both by simulation and by theoretical reasoning, that some spectral model parameters are difficult to estimate using this approach. In contrast, using de-biased Whittle likelihood-based inference, we obtain more accurate and precise parameter estimates. To improve the computational efficiency of de-biased Whittle inference, we introduce a new technique to calculate likelihood derivatives and to approximate the variance of the resulting spectral estimator.