Statistical models for the ocean-atmosphere interaction

The present research aims to model the ocean-atmosphere interaction using the relationship between the Southern Oscillation and El Nino-La Nina events as an input-output system. Statistical modeling of this type of relationship requires analysis of the cross-correlation function between the input and output series. The data considered correspond to monthly time series of the Nino 3.4 Sea Surface Temperature Index and the Southern Oscillation Index between 1982 and 2015. To represent the interaction of the ocean-atmosphere, Function Models were developed Transfer and feedback. The efficiency of these models with data up to 2015 was compared to forecast until September 2019 by means of the mean squared error of prediction, finding that the feedback model is more efficient.