Detection of long-term coherent exchange over spruce forest using wavelet analysis

Summary

This study presents a discussion of a method for automated and quasi-online analysis of coherent structures using wavelet transform. The method is optimised for rapid processing of vector and scalar variables obtained over tall vegetation. It has been designed to assess long-term statistics of coherent structures, as it is applicable over a wide range of atmospheric conditions. Data of artificial and real turbulent signals are used to perform the analysis and to evaluate the presented method.

Different wavelet functions are used for filtering the original signals, determining characteristic time scales, and detecting individual coherent structures. On this basis, statistics of temporal separation of coherent structures and phase shift between different variables can be calculated.

‘Background’ turbulence and spikes are found to be efficiently removed without changing the shape, particularly the sharp localised gradients, of coherent structures. The determined peak in the calculated wavelet variance spectrum is observed to correspond very well to characteristic event durations and to satisfy the definition of coherent structures present in vector and scalar variables. The detection algorithm was successful in analysing data covering a wide range of atmospheric conditions. Detected individual coherent structures provide a parallel temporal pattern for scalar variables, but a phase shift between scalar and vector components.