This present study aims to evaluate the stratospheric ozone that was continuously measured during 2016 over Campo Grande, the capital of South Mato Grosso state, Brazil. To determine the best-adjusted distribution describing the ozone (O₃) co-generation data in Campo Grande, 15 functions were used while modeling the numerical results. Five sets of data were used: the entire year, spring (September to December), summer (December to March, high solar radiation data), autumn (March to June), and winter (June to September, low solar radiation data) to study the seasonal variation in the statistical behavior of the probability distribution functions. The distribution performances are evaluated using three tests of quality, namely Kolmogorov–Smirnov (K-S), Anderson–Darling (A-D), and Chi-square tests. Finally, all the results of the fitted quality tests have been compared. It has been observed that the generalized extreme value distribution provides a good fit all along the year, while for specific seasons the best distributions vary. The best distributions, according to the seasons, are Gamma 3P for the winter, lognormal 3P for spring, Weibull for summer and Gamma 3P for autumn, respectively. There was a coincidence in the probability distribution function adjustment in winter and autumn, period with lower O₃ concentrations, kurtosis, and skewness. In the summer and spring, it was observed higher concentrations of O₃, kurtosis, and asymmetry and different probability distribution functions.

O objetivo desta pesquisa foi de desenvolver um modelo para prever as concentrações de ozônio em função das variáveis climáticas. As medidas de concentração do ozônio foram realizadas na Universidade Federal de Mato Grosso do Sul, utilizando dados do período de janeiro a dezembro de 2014. As variáveis preditoras referentes ao clima (umidade relativa, velocidade do vento, chuva, aeerossóis, índice de claridade, radiação solar global e temperatura) foram cedidas pela CEMTEC-Campo Grande-MS. Foram realizadas análise de componentes principais (PCA) e análise de regressão linear múltipla (MLR) aplicadas a estas séries temporais. As principais relações observadas nestas análises foram então usadas para obter uma equação de regressão que liga as concentrações de ozônio diurnas com parâmetros climáticos. O modelo desenvolvido é capaz de explicar 99% da variabilidade nas concentrações de ozônio de superfície com um erro de 0,2%?