Dinâmica de precipitações extremas em eventos de inundações urbanas : aspectos espaciais e hidrológicos

• Autor(a) principal: Fernando Neves Lima
• Data de Publicação: 2019
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/31416
• topic: Modelagem hidrológica; Drenagem urbana; Inundações bruscas; Engenharia sanitária; Recursos hídricos - Desenvolvimento; Hidrologia; Inundações; Precipitação (Meteorologia)

The urban centers are suffering with flash floods and disordered occupation. These extreme events are becoming more usual. The challenge of hydrology is enormous given the many variables involved: short time response, high catchment’s imperviousness, rainfall spatial and time variability. It is certain that the computational models have advanced in the last years, however the uncertainties are still present in the processes of rainfall-runoff transformation and in the hydraulic propagation of the channels in urban environments. The reduction of these uncertainties was searched in this thesis, both from a hydrological (rainfall input in the models) and a hydraulic (roughness coefficients) point of view, as well as the spatial and hydrological characterization of the intense events occurring in the Arrudas’s catchment located in Minas Gerais, Brazil. At first, a simulation was performed on the SWMM hydrodynamic model. The answer was satisfactory regarding to efficiency.. In a second step, the GR4H concentrated conceptual model was chosen with some methodological adaptations: instead of a traditional calibration, it was proposed to use a list of predefined hydrological model parameters already tested in other basins in line with parameters hydraulics from a set of rating curves. An equifinality was observed, i.e., the hydraulic uncertainties compensated for the hydrological uncertainties. A ranking of parameters was defined based on modeling scenarios and the degree of efficiency. Finally, an innovation from a methodological point of view was held: the stochastic selection and breeding of 12 rainfall events from a geostatistical basis model with the extraction of spatial characteristics (wind direction and speed) and hydrological (intensity and duration) of the rain by select events that presented critical levels in the main channel. This tool has proved very useful as an aid to decision making process regarding urban warnings, especially for the possibility of reducing the lead time alert. The storm tracking made it possible to know features that assist in the flood risk management.