Modelling and forecasting temperature and precipitation in Italy

Abstract

We study the monthly average temperature in Italy for the period 1991-2015. The increase or decrease of the average temperature with respect to the previous year is modelled as a discrete-time Markov chain having four possible states. Similarly, a Markov chain is proposed as a model for the variations of the monthly amount of precipitation. Based on these models, it is possible to forecast whether the temperature and the amount of precipitation are likely to vary significantly in the long term.