ABSTRACT
Cylindrical data are pairs of angles and intensities that are simultaneously observed on the same statistical unit. The name cylindrical is motivated by the special domain of these data because the pair of an angle and an intensity can be described as a point on a cylinder. Correlated cylindrical data arise in environmental studies when cylindrical measurements are repeatedly taken along time or across space, respectively taking the form of a cylindrical time series or a cylindrical spatial series. Popular examples of cylindrical time series include temporal sequences of wind directions and pollutant concentrations [7], wind directions and speeds [18], wildfire sizes and orientations [8], wave directions and heights [4, 17, 24]. Examples of cylindrical spatial series include satellite images of wind directions and speeds [19], wave directions and heights that are generated by deterministic wave models [22, 23], speeds and directions of marine currents recorded by a network of high-frequency radars [17], as well as telemetry data of animal movement [9].
