Data: 17/09/2014 à 19/09/2014
Local: São Paulo - Brazil
Climate Influences on the Flood Hazard Across Brazil (PAP014921)
Impact of climate change on flood risk
Flood risk management and flood frequency analysis have relied on the stationary and independence assumptions of random events, whereas the probability associated with a given flood is constant across time. However, climate variability and land use changes may impose significant modifications in the physical processes associated with extreme rainfall-runoff events in a given region which in turn may alter the flood hazard for short or long periods. In this case, identification of the key climate variables associated with the flood hazard variability is particularly important if one wants to assess the impacts of climate change on the flood risk. Here we explore the hypothesis of a time varying flood hazard for several streamflow gauges in leading locations across Brazil by analyzing historical streamflow series and large scale climate teleconnections associated with their interannual variability. Particularly, we define the flood hazard as the probability of the daily flow to exceed some pre-defined threshold value, e.g., the 90th daily streamflow percentile. For each year and gauge, the number of days in which the daily streamflow exceeds the threshold value is then stored. In order to consider only time independent flood events, we decluster the daily streamflow series by taking only flood events in which the inter-arrival time is greater than three days. The interannual variability in the number of flood events for each site is then associated with climate variability in a Poisson regression framework, whose covariates are representative of ENSO, of the meridional position of the Intertropical Convergence Zone (ITCZ) and of the monsoon system over South America. The spatial distribution of the regression estimates are then used to depict a quantitative view of the most sensitive regions in Brazil in terms of flood risk changes due to ENSO, ITCZ and monsoon variability. Finally, the implications of potential future changes of such large scale climate systems on the flood risk in the country are discussed.