Cities are locations of primary sources for the emission of carbon dioxide, a greenhouse gas thought to be responsible for the majority of recent global warming. Anthropogenic emission sources include fossil fuel burning by industries, for generation of electricity, transport, cooking, etc. These additional emissions modify the natural surface atmosphere carbon exchange dominated by the assimilation/respiration cycle caused by vegetation and soils. A number of studies have shown that carbon dioxide concentrations are on average higher in city centers compared to their rural surroundings. Similarly, the few available direct flux observations also confirm that cities are generally net producers of carbon dioxide. All these studies have exclusively been carried out in cities located in mid-latitude or sub-tropical climate zones. This is unfortunate because much of the anticipated future growth of the world population will occur in cities in the less developed world which are almost exclusively located in the (sub)tropics. This region is characterized by little seasonal variability of atmospheric variables and vegetation development and it is possible that seasonal changes in carbon dioxide concentration and fluxes observed in past studies are not found to the same extend in tropical cities.

Long-term observations of carbon dioxide concentrations carried out by car traverses and using fixed stations located in diverse urban land-uses (instruments included LI8210 and LI840 sensors) and carbon dioxide fluxes from a micrometeorological tower located in a residential area of Singapore (using an open-path IRGA: LI7500) will be presented. Besides a few periods when individual sensors malfunctioned, an almost continuous time series encompassing monsoon and inter-monsoon seasons is available for analysis. The average diurnal carbon dioxide concentration pattern at the rural reference site showed the expected maximum and minimum during nighttime and daytime, respectively. On the other hand urban values peaked during daytime (1330 and 1930 hrs) and experienced a smaller diurnal amplitude (33 ppm compared to 100 ppm at the rural site). At the rural site, the diurnal amplitude was reduced during the relatively wetter NE monsoon period. Unlike in other urban studies, no clear relationship between urban concentration patterns and anthropogenic emissions, such as e.g. from traffic, could be established. The carbon dioxide fluxes showed that the surface-atmosphere exchange on average was positive at the residential observation site despite the relatively large amount of greenspace. This result is consistent with data from other urban studies conducted in mid-latitude cities. Seasonal variability, on the other hand, was clearly absent. Results of turbulent characteristics (integral and spectral) of the carbon dioxide exchange will also be presented if data processing proceeds on schedule.