Inferência e caracterização através de estatística circular para dados de direção do vento

Abstract

Angular data is usually measured in degrees or radians and requires Circular Statistics techniques for characterization and inference. A point in a circle of unit radius or a unit vector indicating direction can be defined as a circular observation. This observation is specified by the angle formed between the initial direction of the circle and the point on the circle corresponding to it, provided that there is a defined initial direction and orientation of the circle. The wind direction is an example that includes angular measurements, which can be given in degrees, with the North cardinal point as the initial direction. The study of wind direction in Brazil can be very useful in the installation of a wind farm, for example. The objective of this study is to characterize the direction of the winds, obtained in degrees to the south of Brazil, comprising the Rio Grande do Sul (RS), Santa Catarina (SC) and Paraná (PR). For this purpose, graphic tools, circular descriptive measures and hypothesis tests were used in order to compare the wind direction in the three regions and relief subdivisions in each state. Daily data were used, collected hourly, made available by the National Institute of Meteorology – INMET. The data were analyzed by state from April 1, 2019 to April 31, 2019.