Estimation of the Envelope Curve for Frequency Factors of Probable Maximum Precipitation (PMP) in Minas Gerais

Código

I-EBHE0121

Autores

Fernando César Soares Feliciano, Eber José de Andrade Pinto

Tema

WG 1.10: Hydrologic Design - Solutions & Communication

Resumo

Extreme precipitation events, such as intense storms and prolonged droughts, are characterized by their rarity in a given hydrometeorological context and can cause substantial material damage and loss of human lives. These events vary according to location and time of year, being influenced by atmospheric circulation processes, topography, latitude, and proximity to oceans. In this context, this study focuses on Minas Gerais, a state in southeastern Brazil, where rainy events typically occur between October and March. The identification of extreme events in hydrology uses deterministic and probabilistic approaches to estimate flows and design hydraulic structures. In the deterministic approach, the Hershfield methodology (1961, 1965) stands out, requiring the mean and standard deviation of a record of annual maximum precipitation. The estimate is made by maximizing the frequency factor k. It is important to note that Hershfield's methodology was primarily developed using pluviometric data from the United States. The climatic and rainfall formation characteristics in the United States differ significantly from those in Brazil. This means that the factors and assumptions used in Hershfield's approach may not accurately reflect the unique meteorological conditions found in Brazil. In this context, this study aims to obtain an envelope curve for the frequency factors "K" of the Probable Maximum Precipitation (PMP) specifically adjusted to the precipitation data from Minas Gerais. The objective is to develop a more representative curve for the state which can contribute to better estimates of PMP for the design and management of hydraulic structures. The first step involved the selection of rain gauge stations in the study region through the open data platform of ANA, applying data quality control criteria. Inconsistent series were analyzed in detail, resulting in the exclusion of inadequate ones. The K factors were calculated for each station following the Hershfield methodology. The envelope curves for the frequency factor were developed using the methodology of Sarkar & Maity (2020). The spatial distribution of the K factors was evaluated by interpolation, using the Inverse Distance Weighting method. The initial filtering identified 2,415 rain gauge stations in Minas Gerais, which were reduced to 484 after applying data quality criteria. The distribution of K factors was right-skewed, with most values between 2 and 4, and three stations showing exceptionally high values. The K values ranged from 1.33 to 13.16, indicating substantial variability. The first quartile (P25%) was 2.41, the median 2.96, and the third quartile (P75%) 3.69. The mean K was 3.27, with a standard deviation of 1.36. These results highlight the regional climatic variability, with daily maximum precipitation averages between 32 and 103 mm. Overall, the results indicate an envelope curve constructed with specific data for the state of Minas Gerais, which presents frequency factors lower than those proposed by Hershfield.