Search for finite dimensional attractors in atmospheric turbulence

The question is investigated whether the dynamics of turbulent wind fields in the atmospheric boundary layer can be satisfactorily described by a low-dimensional deterministic system. Special emphasis is laid on the detection of a possibly existing, underlying strange attractor. Fast response wind measurements of an ultrasonic anemometer with a sampling rate of 21 Hz were carried out over periods of several days in the near surface boundary layer. The correlation dimension of the resulting time series, several million data points long, is estimated by means of the Grassberger-Procaccia algorithm. No sign of a low dimensional attractor can be detected. By comparison with different types of random noise, the existence of an attractor with dimension lower than six can be excluded in the present data sets.