Índice de estresse térmico e sua projeção para cenários futuros no município de Montes Claros – MG

Abstract

Climate change has increasingly compromised human health and well-being, with thermal discomfort becoming an important indicator of these impacts. However, there is still a lack of studies on the subject in cities in northern Minas Gerais. In order to fill this gap, this study investigated how daily and monthly averages of meteorological variables—recorded over the last 25 years in Montes Claros (MG)—influence the THI in the municipality. To this end, time series analysis techniques were used in the R software (version 4.4.1), starting with the imputation of missing values for temperature and relative humidity: missing data from INMET were filled with information from NASA, whose adherence and consistency were assessed through simple linear regressions and several adjustment indices. It was found that NASA estimates present a trend similar to INMET observations, justifying their use as a reliable source for recomposing the series. Next, the THI was calculated at synoptic times (12:00, 18:00 and 24:00 UTC) for the entire historical series and the monthly averages, medians and maximums were determined, on which a preliminary descriptive analysis was performed. To identify trends and seasonality, the Friedman test was applied — with years as blocks and months as treatments — and, subsequently, the SARIMA models were adjusted for future projections. The results confirmed the presence of trend and seasonal components, and indicated the best adjustments to the SARIMA(2,0,1)(0,1,1)₁₂, SARIMA(3,0,2)(1,0,1)₁₂ and SARIMA(2,0,1)(1,0,1)₁₂ models, respectively, for the mean, median and maximum THI. The forecasts indicated an increase of approximately 1unit of THI in the mean temperature over the last two decades, possibly due to the expansion of densely built urban areas, with high impermeability and reduced vegetation cover.