Impacts of climate oscillation on precipitation and river flows of small scale river basins in Eastern Northeast Brazil

Sammanfattning: The Northeast region of Brazil is one of the most susceptible parts of the country to the effects of climate variability leading to severe droughts or extreme floods. Indeed, drought is a major issue due to a large semiarid area that covers 53% of the region, which determines the main climate characteristics of this region. On the other hand, extreme floods have caused widespread damage in river basins in the eastern part of Northeast. The Eastern Northeast is the most humid area of the Northeast region. It comprises the Atlantic rain forest and may receive up to 2000 mm precipitation a year, but 70% of total falls are concentrated from April to August, the period defined as its rainy season. The remaining of the year, when 30% of the yearly precipitation falls, is considered as the dry season of the region.This study focuses on investigating the spatial and temporal variability of the seasonal precipitation at river basin scales in Eastern Northeast Brazil, and how climate oscillation factors are connected to the inter-annual variability of the precipitation over the region. This study has shown that the local rainfall variability is dominated by a multi-annual variation related to the tropical Atlantic Ocean conditions and intensified by the tropical Pacific.Furthermore, flood events were analysed in one of the most affected river basins in the Eastern Northeast region, the Mundaú river basin. The extreme flood event of June 2010 was reproduced by combining hydrological and hydraulic modelling and probabilistic flow estimations. Due to the lack of recorded flows during the event, secondary data sources were used such as water-level to estimate flows from the stage-discharge relationship, and post-event flood marks to calibrate a hydraulic model. Thus, it was possible to simulate the flooded areas of the June 2010 floods, and mapping areas susceptible to flooding for different return periods.The hydrological model was also used to explore flood risk monitoring by using high-resolution data from automatic rain/discharge gauge stations and weather radar, which has recently been installed to monitor river basins of Eastern Northeast Brazil. The radar capability on estimating rainfall was qualitative and quantitative evaluated, and combined with observed rainfall to be applied as input on the hydrological model as a basis for the development of a tool for flooding warning system.