Turbulent exchange of carbon dioxide in a complex urban environment : results from Long-term Eddy Covariance measurements

Within this thesis, characteristics of the turbulent exchange within the urban boundary layer, as well as long-term trends and tendencies of the carbon dioxide flux (FC) and concentration (rhoC) are presented. Prevailing transport processes, the transfer and transport efficiencies of the turbulent exchange of momentum, heat, CO2 and H2O and the importance of coherent turbulent motions within the urban boundary layer are studied by applying the quadrant analysis technique. The behavior of FC and rhoC in the urban environment is investigated on daily, seasonal, and inter-annual scales as well as in comparison to regional background concentration records of the atmospheric CO2 concentration.
The dependence of the passive scalars CO2 and H2O on atmospheric stability (zeta) differs distinctly in comparison to momentum and heat. The vertical fluxes of momentum (tau) and heat (QH) are actively generating mechanically and buoyantly driven turbulence, respectively. Due to the strong coupling between tau, QH and zeta, each stability class is characterized by a distinctive turbulence regime. In contrast, the turbulent exchange of CO2 and H2O is not primarily controlled by the existence of transporting eddies, but also heavily influenced by the activity and the composition of the corresponding scalar sources (e.g. traffic, heating etc.) and thus, the heterogeneity of the local surrounding. Models represent the transfer efficiencies of momentum and heat accurately while the prediction for CO2 and H2O mostly fails. Other factors, like the interplay between the activity of sources and sinks are more important and accordingly, the transfer efficiency of CO2 can be consulted to identify times or wind sectors where the source/sink regime is altered by e.g. photosynthetic activity. The inter-comparison of the transport characteristics of heat, CO2 and H2O leads to the assumption of scalar dissimilarity. By applying the quadrant analysis framework to the long-term time series, dominant turbulent structures responsible for efficient vertical exchange, i.e. coherent structures, can be identified. The length of the time series allows to extend the analysis to the stable range, which usually rarely occurs in urban areas.
The variability of local urban rhoC is investigated in comparison to regional background concentration records. While patterns on daily and seasonal scales are similar, the vicinity to the ground sources of the local measurements leads to a stronger sensitivity to changes on small temporal scales. The height above ground of the background concentration measurements and thus the larger distance from the ground sources results in a phase shift of up to three months compared to the local seasonal course of rhoC. While rhoC in the urban area is also clearly elevated by 10 ppm on average, the behavior of rhoC in the urban environment reveals good consistency with background concentration measurements in terms of seasonality and long-term trend. The calculated local linear trend for the time period between 2005 and 2014 is around 2 ppm/y, which also coincides well with the global average trend.
The coupling between FC and anthropogenic activity in the urban area is apparent from considerable differences in weekday and weekend fluxes, the diurnal cycle as a result of traffic volume or the seasonality caused by additional heating activity in wintertime. The variability of FC scales with the source activity and a long-term decrease of FC around 5% is observed locally as a result of a decrease in traffic activity during the investigation period. However, variabilities on all temporal scales are clearly larger than the observed long-term tendencies. For the investigation of FC in the heterogeneous urban environment an appropriate weighting between individual wind sectors is shown to be necessary due to the unequal frequency distribution of wind directions. The application of a refined methodology for the calculation of horizontally averaged fluxes of CO2 significantly improves the representativity of the data for the investigation area and also enhances the comparability of the data to results from other studies. The length of the current dataset allows to estimate the significance of the observed long-term behavior. While up to six years are potentially needed to calculate a significant inter-annual trend of rhoC, statistics of FC still benefit from even longer data records due to the larger variability. This gives evidence, that long-term time series of urban CO2 can help to add valuable knowledge to the current understanding of the urban ecosystem and its role in the global carbon balance.