Carbon and water exchange in extensive green roofs: A comparison between eddy covariance and soil flux chamber observations

Vegetated roofs are currently discussed as a mitigation strategy to limit the impact of climate change in urban areas by sequestering carbon dioxide (CO₂) via photosynthetic uptake. Two observational methods, i.e., eddy covariance (EC) and automated soil flux chambers, were co-located on a large extensive green roof in Berlin, Germany, to quantify the surface-atmosphere exchange of carbon and water over a period of 4.5 months. The relationship between the surface chambers and the eddy covariance technique (EC) was poor for water vapour exchange, with the chambers underestimating half-hourly EC fluxes by up to 75 %. The analysis showed a good overall agreement between the EC carbon fluxes and the chamber-averaged fluxes on a half-hourly basis. However, the net carbon uptake of −14 g C m⁻² over the study period as observed by EC, was overestimated by the surface chambers which showed an uptake of −51 g C m⁻². The deviation is attributed to the spatial heterogeneity of the green roof site, e.g., individual plant species, and to the microscale variation of substrate properties, which is more strongly detected by enclosed chambers located on small source areas, in contrast to eddy covariance observations, which integrate a much larger source area. However, soil flux chambers were found to be an adequate alternative to EC measurements for carbon exchange, given that multiple chamber measurements spots are selected to integrate and assess the spatial heterogeneity on green roofs.