Community structure detection in complex networks for characterizing atmospheric boundary-layer wind speed time series

Better understanding of boundary-layer wind field is fundamental for modeling of various wind-related phenomena, e.g., gas dispersal and energy generation. In this paper, we first record three groups of the near surface wind speed time series in different locations using a high-resolution 3D ultrasonic anemometer. Then, we map these experimental time series into complex networks and detect the corresponding communities based on the modularity analysis. Next, we extract two network characteristics, e.g., clustering coefficient and average shortest path. Finally, we find that although the wind speed time series are recorded at different locations, the fluctuation behaviors exhibit the similar community structures. This interesting finding suggests that network community structure analysis allows better understanding of the inherent structures of nonlinear and nonstationary near-surface wind speed signals.