Tuning of ARMA models for mono-modal sea spectrum estimation

The paper focuses on the tuning of Auto Regressive Moving Average (ARMA) models aimed at describing the spectral behaviour of sea wave elevation time series. The analysis considers mono-modal sea states, mainly due to the wind effect. The main problem related to the estimation of such models, starting from the wave elevation time series, concerns the tuning of the parameters involved in the identification algorithms. Indeed, the value of some parameters must be chosen by the user and these choices strongly affect the goodness of the estimates. Here, some of these parameters are considered, highlighting their effect on the goodness of the fit, and indications about how to set their values for applications related to mono-modal sea states are presented. The results are fully discussed considering a reference case study and a procedure to tune the model is proposed.