The Application of Informational Predictability to Rainfall Data

Informational predictability, as defined in Fernández Méndez et al. [Stoch. Environ. Res. Risk Assess. (2023), submitted for publication] is based on the normalized complement of the expected value of the logarithm of the conditional probability, to be precise, this refers to the probability of the predicted events, when conditioned upon their respective predictors. The present work focuses on balancing the precision of the prediction, as measured by the narrowness of the predicted intervals, against the respective probabilities of a correct prediction, which finally amounts to maximizing the informational predictability. The data are high-resolution temporal rainfall intensity series, measured by an optical rain gauge.