Investigating the near-surface airflow and its turbulent and submesoscale statistics for the weak-wind regime from field experiments
Anita Freundorfer (11/2016-11/2017)
Support: Christoph Thomas, Ingo Rehberg
The behaviour of the atmospheric boundary layer (ABL) during weak-wind situa- tions is still not completely understood, mainly due to the existence of so-called submeso motions. Thus, further research on the typical phenomena during weak-wind situations and their physical mechanisms is required. For this purpose, during the Advanced Resolution Canopy FLow Observations experiment (ARCFLO), data were collected at four different sites located in Oregon, USA. Those sites cover a variety of terrain complexities and vegetation densities, especially focusing on different forest architectures. The following work focusses on evaluating the data obtained in the series of ARCFLO experiments and comparing the different sites. After an introduction to the typical approaches for investigating turbulence in the ABL and a description of the study sites and our data, a way of determining an objective threshold velocity for the weak-wind regime is elaborated. This is done using a scatter plot of the friction velocity u ∗ in dependence of the mean wind velocity u skal . Using a segmented linear regression with two segments, a transition point can be found, where the mean behaviour of the friction velocity changes from nearly independent from the mean wind velocity to a linear dependency. This threshold can be identified for all four sites. The identified thresholds vary from u thr = 0.25 ms at the grassland site to u thr = 1.03 ms most densely forested site. Subsequently, it is investigated how the weak-wind threshold is influenced by the landscape and vegetation. We find that a denser vegetation leads to a higher threshold velocity. Thus, we test the hypothesis that in the weak-wind regime, the subcanopy is inherently decoupled from the above canopy. However, this hypothesis can not be confirmed. In the next section, meandering is analysed. Meandering is one of the typical phe- nomena during weak-wind situations. We develop a novel method for identifying meandering periods using the difference between vector averages and scalar averages. The results from this new method are compared to those from a commonly applied method from literature. This other method makes use of the special form of the autocorrelation function of the horizontal wind components in meandering situations. For the detected meandering periods we analyse the typical inherent time scales by means of the autocorrelation function as well as by using wavelet analysis. We find a broad range of meandering time scales at all 4 sites with a preference of shorter time scales. The average time scale during the night is slightly longer than that during the day.