Non parametric modelling of cyclo-stationary markovian processes - Part I: Simulation of multivariate sea state processes

The study of a large range of Marine dynamical systems such as structure reliability, transport of sediment, erosion, etc. requires to understand the long term time evolution of sea state processes. We address in this paper the non parametric modelling of cyclo-stationary multivariate Markov processes using a continuous state space and discrete time Markov chain for which all necessary densities are approximated non parametrically using observed samples. This new resampling technique, is tested on hindcast data in the estimation of mean time necessary to perform an offshore operation given the date of the beginning of the operation.