Metodologia para consideração de tendências climáticas em curvas IDF de precipitação

• Autor(a) principal: Tamiosso, Marilia Ferreira
• Data de Publicação: 2019
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Manancial - Repositório Digital da UFSM
• Texto Completo: http://repositorio.ufsm.br/handle/1/19212
• topic: Tendências climáticas; Chuvas extremas; Drenagem urbana; Climate trends; Extreme rainfall; Urban drainage; CNPQ::ENGENHARIAS::ENGENHARIA CIVIL

Many studies show that extreme rainfall have changed their behavior in recent years, caused by climate change. This fact has consequences such as increased occurrence of flood and other critical events in urban areas. This study aims to contribute to the analysis of extreme events trends and how these can be considered in the hydraulic projects, seeking to reduce the occurrence of urban disasters due to undersizing. For this, synthetic series of annual maximum rainfall with duration of 1 hour and different magnitudes of trend were created. It was analyzed the required size of observed series for the detection of trends and the impact of trends in hydraulic projects. Two methodologies were also tested to update the IDF curves in case of trend detection in the annual maximum rainfall series. The first analysis showed that is required a serie with about 98 years of observed data to detect trend magnitudes (with α = 0,1 of significance) equal to 0,05mm/year, while a serie data about 27 years is required to detect magnitudes of 0,40mm/ano. The results showed how uncertain is the trend detection, the smaller the series size. In relation to the trends impacts in hydraulic projects, it was possible to observe that for some specific situations with mild magnitudes, the flow increment for future situation is quite small, and may even be neglected, while in other cases, the future flow increase situations poses great risk if the pipeline works undersized over the years. The methodologies for considering the trends obtained acceptable performance for series with size from 50 years of observed data, for return periods with up to 100 years and lifetime up to 50 years. For smaller series, the methodologies presented extremely large errors due to the low representativeness of the short series, being the magnitude detected at the beginning of the series different from that detected at the end of the lifetime of the structure. It is noticed that the rainfall data limitation in our country brings us to a situation of great uncertainty regarding the trends of maximum rainfall series, where most series have sizes with about 10 to 20 years of observed data.