Caracterização morfométrica da bacia hidrográfica do Ribeirão Vermelho e sua relação com índices de exposição ao vento e desastres

Abstract

Wind is a key climatic variable and is of interest in various applications, including infrastruc- ture damage assessments, environmental and socio-economic risks, and is important in wind energy development. It is considered a meteorological element that is difficult to predict due to atmospheric dynamics and the impossibility of providing an accurate forecast with traditional statistical forecasting models. The aim of this study was to characterize the spatio-temporal variation in wind speed, direction, trend and disasters caused by wind along the Ribeirão Ver- melho watershed in the state of Minas Gerais, Brazil. We used average daily wind speed and direction data from the historical series 2010 to 2023 and monthly data for the year 2024, col- lected by a universal anemometer located 10 m above the surface, belonging to the National Institute of Meteorology (INMET), available through the Meteorological Database for Educa- tion and Research (BDMEP).Using the R computer language, basic statistical tests and wind direction frequency calculations were carried out, generating tables and figures to facilitate cha- racterization. SARIMA and SARIMAX models were used to forecast wind speeds for 2025. The exposure and protection indices were also determined using geoprocessing techniques in R, using the digital elevation model, wind speed (1.83 ms−1 ) and the predominant direction (east) to model the wind redistribution parameters in the region. The average and the Mann-Kendall statistical test showed a reduction in wind speed. The average wind speed for the 2010-2023 series was 1.83 ms−1 , with the lowest monthly average of 1.54 ms−1 recorded in June and the highest of 2.11 ms−1 in September. There was a significant reduction in wind speed in the rainy season with strong instability, the most predominant direction being east (31.8%). The study region has locations of low, moderate, high and very high risk of disasters caused by winds, depending on the variation in relief. However, in 2022 and 2023 there was a higher coefficient of variation in wind speed (100.0%), indicating that the wind has become more variable and less predictable, increasing the risk of disasters caused by winds equal to or greater than 5.54 ms−1 as occurred on September 27, 2023 affecting the urban region of Lavras, belonging to the Ribeirão Vermelho watershed. Around 0.11 km2 is at very high risk and 13.57 km2 is at high risk of disaster, especially during the rainy season. The SARIMAX wind forecasting model is indicated as being a preferred model for wind forecasting over SARIMA, due to its efficiency and robustness in adjusting to the original series and it also showed lower forecast errors.