Representativity of CO ₂ Exchange on Extensive Green Roofs: Scaling From Leaf to Plant, and Plant to Roof Scale

Abstract

Urban green roofs have potential for local urban climate change mitigation, as they have been demonstrated to uptake carbon dioxide (CO ₂ ) from the atmosphere. However, most roofs are too small to qualify for area‐integrating direct surface‐atmosphere exchange measurements using the eddy covariance technique (EC), resulting in sparse data on net CO ₂ sequestration capacities. Small‐scale methods such as leaf chambers or automated surface chambers (AC) may provide a monitoring alternative for smaller roofs. To assess whether alternative methods can produce representative data on the CO ₂ sequestration capacity of an extensive green roof, simultaneous measurements at the scale of leaf, plants, and roof were conducted. The representativity of leaf scale measurements was strongly affected by spatial heterogeneity and high temperatures within the chamber. Linear regression indicated accurate scaling from plant to roof scale (slope = 0.76) at diurnal mean temperatures around 14.2°C and recurring precipitation. However, with daytime temperatures exceeding 25°C, AC more significantly underestimated the roof CO ₂ assimilation capacity in comparison to EC (slope = 0.63). Data from a prolonged drought period indicated that certain small‐scale responses to water stress, such as the shift to CAM photosynthesis, captured by AC are not registered by EC. A spatial upscaling based on the estimated vegetation distribution on the roof indicated a limited consistency with the EC data with slopes of 0.20 and 0.42 for the aforementioned periods. Differences between AC and EC should not necessarily be regarded as errors but as valuable insights into spatial and plant‐physiological heterogeneity.