Por Artur Tiago Silva (CERIS, Instituto Superior Técnico, Universidade de Lisboa).

Due to the increasing consensus on the impacts of climate change and variability on the frequency of hydrological extremes, scientific advancements on dealing with hydrological change are required. The objective of this work is to develop and build upon existing analysis frameworks for modeling floods under nonstationarity while critically assessing the implications of these techniques for hydrological science and practice and, concomitantly, to develop research on possible climate-flood links.  The research is supported by two main case studies. In the first one, an exploratory analysis of flood occurrence rates in 10 Portuguese catchments reveals that the years with fewer floods tend to happen when the winter North Atlantic Oscillation index is in its positive phase and vice-versa. The second case study focuses on the Itajaí-açu River in Southern Brazil, which has a long history of extreme floods. Results show statistically significant and complex dependence relationship between the flood regime and the El Niño - Southern Oscillation (ENSO).