Nadarajah Ramesh, Gayatri Rode
Hidden Markov models provide a rich class of stochastic models that are very useful in hydrological studies. This paper describes a class of hidden Markov models that incorporate covariates in their state distributions to model daily rainfall time series. Greater emphasis is placed on nding a model that can reproduce the second-order properties of the observed rainfall sequences. We present the construction of the likelihood function incorporating time-dependent atmospheric covariates in rainfall distributions. The performance of the model is assessed using daily rainfall data from Leicester, East Midlands region of England.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados