Multiscale entropy analysis of the 3D near-surface wind field

Characterizing nonlinear dynamic behaviors in the 3D near-surface wind field is a challenging problem of significant importance. The study of pollutant transport, wind power evaluation, forest damage assessment, among others, would benefit from this research. In this paper, we first use a high-resolution 3D ultrasonic anemometer to simultaneously measure three wind components. Then the multiscale entropy (MSE) and multivariate multiscale entropy (MMSE) method are employed to quantitatively analyze complexity of the 3D wind speed/direction signals with different time lags respectively. Results suggest that the vertical wind speed time series is more complex than the two horizontal ones. Moreover, the results of MSE show that the complexity of one vertical wind direction signal is similar to that of the horizontal one, while the complexity of another vertical wind direction signal is very low compared with the horizontal one. Furthermore, the results of MMSE indicate that the two combined vertical wind direction signals and the horizontal one exhibit very similar complexity. These interesting findings suggest that the MSE and MMSE can potentially be useful tools for uncovering the complex dynamic behaviors underlying the 3D wind variation records.