Pluvial Flooding in Santo André City - São Paulo: Observation and Prediction (PAP014774)

Código

PAP014774

Autores

María Cleofé Valverde Brambila, C. L. dos Santos

Tema

Urban Floods

Resumo

The goal of this paper is to analyze the pluvial flooding that occurs in Santo André region caused by extreme rainfall as well as to propose a statistical model for rainfall flood forecasting. Santo André being located in the São Paulo metropolitan area, it has developed in the floodplains of the Tamanduateí River, and for this reason it has a history of recurrent floods. In this study, three points of the highest frequency of flooding were selected: the river Tamanduateí and its tributaries Meninos and Oratório. The dataset includes daily rainfall data of three rain gauges closest to the rivers. So we carried out a statistical analysis involving the accumulated daily rainfall with flood records on the three rivers. On the other hand, multi-linear regression (MLR) approach is used to construct the daily extreme rainfall prediction model. The results showed that the highest frequency of rains and flooding occurs in the months of January and February with an average of 2 flood events in January. The floods identified as extreme by the number of affected districts occurred in 15 December 2011, where the rainfall intensity exceeded the daily value of 120 mm. On the other hand, the higher frequency of flood events was associated with the overflow of the river Tamanduateí followed by that of Meninos. The maximum daily rainfall associated with these events reached values of 152 mm. In addition, it was confirmed that the rains, both in their mean values, as in their extremes, have increased. The MLR model showed difficulty in capturing the peaks of extreme rainfall that cause floods, underestimating their value. We believe that the MLR model cannot follow the inherent non-linearity of rain. However, the best prediction model was found for the river Meninos, where the model followed the variability of daily rainfall.