Avaliação da influência de variáveis naturais na temperatura do ar através de regressão linear

Evaluation of the influence of natural variables on air temperature through linear regression

Abstract

Air temperature is a factor that directly affects the physiological processes of plants and animals. Therefore, its understanding has become essential in analyzing the adaptations of fauna and flora in certain areas, as well as in studies related to human thermal comfort. However, a major obstacle to this territorial characterization is the lack of a dense network for monitoring climatic data. Various methods have been applied to estimate temperatures in areas without meteorological data, including multiple linear regression equations. This study aims to create a model to predict the average air temperature based on altitude, latitude, longitude, and surface temperature obtained from orbital sensors. The choice of the study area encompassed the states of the South, Southeast, and Mato Grosso do Sul, with 319 weather stations. The resulting regression model achieved an R² of 0.74. Temperature predictions and residuals were generated, highlighting higher temperatures in the Midwest, demonstrating the influence of the continentality phenomenon, and lower temperatures in mountainous areas due to altitude. This study underscores the importance of climate monitoring in understanding geographical influences on air temperatures.