Search

Floods and droughts are natural phenomena of difficult prevision. Awareness, mitigation and water resource management, depend upon timely information on the beginning of a dry or wet period and on their extension in time and space. This study predicts SPI steady state class probabilities in the Douro region, Portugal, through an empirical and Markov chain approach, expected residence time in each class of severity; expected first passage time, recurrence time and class prediction in a short-term basis. It was verified that all these variables are useful for water resources management. [ABSTRACT FROM AUTHOR]