An autoregressive model with time-varying coefficients for wind fields

In this article, an original Markov-switching autoregressive model is proposed to describe the space–time evolution of wind fields. At first, a non-observable process is introduced in order to model the motion of the meteorological structures. Then, conditionally to this process, the evolution of the wind fields is described using autoregressive models with time-varying coefficients. The proposed model is calibrated and validated on data in the North Atlantic.