Testing the Need for Replication of Eddy Covariance Carbon Dioxide Flux Measurements over Agricultural Fields

Abstract:

The eddy covariance method directly measures carbon dioxide (CO₂) fluxes for long periods of time and with footprints up to hundreds of meters in size. Any ecosystem process that alters how gases and energy move between the atmosphere and soil/vegetation can affect these fluxes. Eddy covariance is vulnerable to systematic errors and uncertainy, particular through relying on assumptions about surface characteristics. Additionally, spatial variation within a site can cause more uncertainty in these measurements and lack of replication in many eddy covariance studies makes statistical analysis of carbon fluxes challenging. We tested if there are significant differences between co-located and simultaneous CO₂ flux measurements over a uniform crop surface, and if the differences increase if we measure different flux footprint areas over the same field. During the summer of 2014, three matched instrumented 2.5-m high towers were co-located and then periodically separated by moving at 50 m intervals along a north-south transect on an alfalfa/trefoil field and a spring wheat field in Southern Manitoba, Canada to compare CO­₂ fluxes. Georeferenced leaf area index measurements were taken in 50 m grid of each field to establish uniformity of the source/sink within a footprint. Diurnal differences of similar magnitude in the CO_{2 ­}fluxes were found in both the co-located experiment and the spatially separated intervals. Despite rigorous calibration during the experiment, some differences were caused by the measurement systems rather than by variation within the field. Interpretation of the spatial variation in leaf area index is being used to determine the contribution caused by difference in source/sink contributions to the flux footprint areas when the towers were spatially separated.