Abstract

Log-linear quasi-association models have been successfully applied to analyse and predict drought class transitions derived from standardized precipitation evapotranspiration index (SPEI) time series in Portugal. This kind of models proved to be suitable for fitting the SPEI drought transitions and is considered a reliable tool for capturing the dynamics of drought severity changes since it estimates the probabilities associated with transitions between the different drought severities. In the context of groundwater drought monitoring, the standardized groundwater index (SGI) is used and is computed from groundwater levels available from the SNIRH piezometric network. The aim is to employ similar models to the transitions between SGI drought classes and compare them with the SPEIs of different accumulation periods for Algarve region in southern Portugal. The purpose is also to evaluate the effectiveness of these tools in predicting short-term transitions in groundwater drought. The findings contribute to improving water management practices and enhancing early warning systems to mitigate the impacts of drought in the Algarve, with potential applications in other parts of the world.