Impacts of climate change on the intensity-duration-frequency curves of two urbanized areas in Brazil using the high-resolution Eta atmospheric model.

In Brazil, little is known about climate change impacts on extreme events at local scale. Here, we explored ways of communicating it, in the form of intensity-duration-frequency (IDF) curves, using the nonstationary equidistant quantile matching (EQMNS) method. EQMNS was applied in two urbanized areas of Brazil: São Paulo and Porto Alegre. This research was conducted using recorded and modeled precipitation data, the latter using the Regional Climate Model Eta. Results indicated an increase in rainfall intensity for rare events, which were intensified under nonstationarity. The use of stationarity misrepresented the trends related to precipitation, being able to mask the direction of change in future conditions. Future IDFs with Eta's 3-hourly resolution appeared to be intensified when compared with those from daily resolution. The increase in rainfall intensity was more significant for short durations than for daily ones, bringing concerns about associated hazards and their impacts on society and urban infrastructure. [ABSTRACT FROM AUTHOR]