Aplicações de análises multivariadas às variáveis que compõem o índice mineiro de vulnerabilidade climática

Abstract

This study aims to interpret the variables of the Mineiro Climate Vulnerability Index (IMVC) through factor analysis, seeking to deepen the understanding of climate vulnerabilities in the municipalities of Minas Gerais. The variables were standardized and extracted from the CLIMA GERAIS website to ensure the effectiveness of the analyses. Confirmatory Factor Analysis (CFA) was employed to assess the adequacy of the original IMVC model, while Exploratory Factor Analysis (EFA) revealed latent factors that explain a significant portion of the data variance by grouping correlated variables. EFA managed to reduce the dimensionality to four main components, which explain a considerable part of the total variance, suggesting that these components efficiently represent the structure of the original data. Additionally, cluster analysis was used to identify vulnerability patterns among municipalities, using five groups to classify according to the level of vulnerability of the original index. However, a significant discrepancy was observed between the results of the statistical methods and the initial classification suggested by the IMVC. The results were validated through Spearman and Kendall correlations. This study highlights the relevance of applying statistical methods to identify vulnerability patterns in Minas Gerais and the need to revise the classification criteria of the index, aiming to optimize public policies for climate adaptation and risk management.