Timescales and structures in vertical transport in the atmospheric boundary layer

A combination of high frequency measurements from the Cabauw measurement site and LES results is used to evaluate timescales and structures of turbulent transport of heat in the atmospheric boundary layer. On the basis of the cospectrum of vertical velocity and temperature, a dominant timescale of transport is introduced. An emperical equation is proposed to easily determine this timescale as a function of wind velocity, height and Obukhov length. Using this timescale, loss in Eddy Covariance measurements due to finite averaging time can be predicted on the basis of a model for the cospectrum shape. Eddy Covariance measurements from the Cabauw site show behaviour confirming this predicted loss behaviour. No evidence is found for the possible presence of unpredicted systematic flux on very large timescales. Attention is also given to the presence of a certain gap scale seperating turbulent and mesoscale motions, but it is found that these motions overlap in unstable conditions, making seperation impossible. As several authors have pointed at the phenomena of Turbulent Organized Structures as large contributors to large timescale flux, an attempt is made to identify these structures in measurement series. This is done with the help of two algorithms, based on criteria and wavelet analysis, demanding vertical coherence and rising, warm air. Detected structures within measured and LES simulated time series show great similarity, giving further confidence that detection is possible and appropiate. Thermals are found as coherent structures of warm and rising air, surrounded by cold and descending air. A statistical profile is made of the shape of thermals as they appear in time-series, which is confirmed by LES results. Behaviour of the thermals involving size, velocity and temperature are investigated for different heights. The average flux within thermals is found to be well modelled by the mass flux approach when based on a thermal surface fraction of 10 percent.