| Popis: |
The notion of decomposing time series of winds and scalars into a mean and a fluctuating component, historically known as Reynolds decomposition, is essential for the determination of the turbulent fluxes in the atmospheric boundary layer. Proper estimates of these fluxes are crucial for a number of applications, including calculating the annual budget of the net ecosystem exchange or the development of new turbulence closure parameterizations for numerical weather prediction. Assuming that the turbulent, microscale motions of interest are separated from macroscale motions by a spectral gap, one can readily apply Reynolds decomposition. Several studies have reported a range of scales that fall into such a spectral gap, especially for flat terrain. However, spatial and temporal information of the spectral gap scales are still lacking, especially over complex terrain. Micro- and mesoscale flows in complex terrain, such as valley and slope flows, may degrade the presence of a spectral gap and/or shift it to scales that are different than those over flat terrain. Thus, more information is needed about the variability of gap scales over complex terrain. The main dataset we analyze is comprised of measurements obtained during the Terrain-Induced Rotor Experiment, conducted in the spring 2006 in Owens Valley, CA. Specifically, we analyze time series of winds obtained on 16 weather stations spread out across the valley, as well as sonic anemometer measurements of winds and temperature at two locations. To these we apply two methods to determine the climatology of gap scales, including the fast Fourier transform and multiresolution flux decomposition. We also focus on several phenomena typical for complex terrain, such as rotors and upslope flows, with the goal of determining if and how gap scales react to their occurence. Our results indicate that the typical range of gap scales is between 17 and 29 min, with substantial cross-valley variability and along-valley homogeneity. At the high-frequency end of the spectral gap, rotors and the deepening of the convective boundary layer are found to be the main gap scale drivers. This work has also gained insight into the low-frequency end of the gap, where the valley-slope flow system was found to be the dominant phenomenon. Here, the dominant mode of upslope flow variability, ranging from 80 to 200 min, was both observed and modelled using a simple periodicity model. |
