Variability of carbon dioxide fluxes in heterogeneous urban environments : from street canyon to neighborhood scale

High carbon dioxide (CO2) emissions in cities are a consequence of high energy consumption by a dense population. Fossil fuel related emissions from urban areas contribute to the increase in average atmospheric concentrations of CO2. Typically, the size of the emissions has been estimated through indirect approaches on the base of energy consumption data, but the number of direct atmospheric measurements of the effective CO2 exchange over cities is increasing since about a decade. Results so far show a lot of uncertainties concerning the processes that control the exchange rates to the atmosphere, i.e. the vertical CO2 flux (FC). The wide range of FC reported for cities reflects the variety and complexity of urban areas. It shows that the urban structure around an observational site has a great influence on the measured flux and that additional controlling factors need to be taken into account when addressing urban flux patterns. In this thesis, the main controlling factors for the variability of FC on different scales in time and space are identified. Long term CO2 concentrations and fluxes are observed at two urban observational sites to account for spatial differences within a city. Micro to local scale exchange processes and spatial distribution patterns between individual urban structure elements are addressed by additional measurements inside a street canyon. The key factors that control FC in heterogeneous urban areas are the major emitters of CO2 and their typical cycles on different time scales. Traffic emissions account for average diurnal courses of measured fluxes, while heating related emissions explain seasonality. Sink effects related to photosynthesis are found to be negligible in most cases. The spatial proximity of major roads at both sites is of greater importance than the source area extent and traffic induced FC is estimated to account on average for 70% of the total flux at one of the sites. Wind direction is - in combination with the spatial distribution of sources - the third and most crucial controlling factor due to its typical cycles and its occasional variability. Diurnal patterns interact with the course of traffic and seasonal characteristics with heating emissions. Both account for typical site specific FC patterns, whereas deviations from usually prevailing wind directions can lead to flux variance of noticeable size. Wind directions are also responsible for micro scale distribution and exchange patterns in and above the street canyon. High horizontal concentration differences inside the canyon are a result of in-canyon flows in the form of a helical vortex. Depending on the prevailing wind direction, vertical exchange of CO2 with the layers above is restricted or enhanced. Annual carbon exchange rate is a typical unit used for comparisons between sites and cities. The spatial heterogeneity of emissions induces a bias into these exchange rates that is related to an unequal reflection by wind direction frequencies. Up-scaling of spatially segregated fluxes leads to a comprehensive and more representative total carbon exchange rate of the average source area than FC does. The identified controlling factors explain the patterns of measured FC and the presented results improve the understanding of the variability of carbon dioxide exchange rates in urban areas. City scale differences and the strong relation on the interdependency of the controlling factors and their typical cycles suggest that the reasons for varying FC in cities worldwide are related to similarly complex local effects