Geocartografia e métodos estatísticos aplicados à análise espaço-temporal de tendência dos extremos pluviométricos em Minas Gerais - Brasil

Abstract

Climate change has significantly impacted the frequency and intensity of extreme precipitation events, posing a challenge for regions such as the state of Minas Gerais, Brazil. In this thesis, a detailed and methodologically rigorous analysis was conducted to understand these impacts, using raw precipitation data from 232 rain gauge stations distributed across Minas Gerais over 44 years (1977-2021), without data imputation. The decision to work with raw data ensured the integrity and authenticity of the collected information, allowing for an accurate and reliable analysis of precipitation extremes and their patterns over time. Unlike other approaches that analyze seasonal data, this thesis adopted a monthly approach, enabling the identification of seasonal patterns and trends specific to each month. This approach captured more subtle and detailed temporal variations, offering insights into how extreme events occur throughout the year, highlighting nuances that could be overlooked in broader seasonal analyses. The application of statistical downscaling was a methodological highlight, allowing the adaptation of global climate projections to a more local and detailed scale. This technique was essential for understanding how climate change influences extreme precipitation patterns in different regions of Minas Gerais, providing an accurate view of precipitation variability and extremes at the regional level. By using data from local rain gauge stations, the study was able to identify patterns that might be masked in analyses with lower resolution, offering a more complete picture of the distribution of extreme events in the state. The results showed a trend of decreasing precipitation extremes during January across the state, while the months of February, March, October, November, and December showed a significant increase in such events. Spatial analysis revealed that the central region of Minas Gerais exhibits greater stability in extreme precipitation patterns, while the eastern and northeastern regions demonstrated greater variability and trends of decreasing extremes. These findings reflect the influence of local and global climate processes, indicating that climate change is differentially impacting the distribution and intensity of extreme events across the state. The use of statistical tests such as ADF (Augmented Dickey-Fuller), KPSS (Kwiatkowski-Phillips-Schmidt-Shin), Mann-Kendall, and Modified Mann-Kendall allowed for the identification and confirmation of trends and non-stationary patterns in precipitation time series. The combination of these tests strengthened the robustness of the analysis, ensuring the precise identification of trends associated with climate change. This research is particularly relevant in the context of climate change, as it offers a detailed understanding of the variability and trends of extreme precipitation events at the local scale. The decomposition of the time series from some stations with significant trends, both of increase and decrease, especially those that showed non-stationary behaviors, revealed that there has been a shift in local precipitation patterns. This suggests that such changes may be driven by seasonal climate phenomena, but mainly by global climate change. The results are crucial for improving regional climate predictions and developing more localized adaptation and mitigation strategies in Minas Gerais. By providing detailed information on the influence of climate change on extreme precipitation patterns, the thesis makes a significant contribution to the development of public policies aimed at water resource management, urban planning, agriculture, and natural disaster prevention. In conclusion, the research represents a significant contribution to understanding climate variability in Minas Gerais and underscores the need to use raw data, monthly analyses, and techniques such as statistical downscaling to understand the impacts of climate change at the regional level. The information generated by this thesis offers valuable insights to enhance local climate forecasts and address the challenges associated with extreme precipitation events in the state of Minas Gerais.