LAUSR

Abstract The net ecosystem exchange (NEE) is the difference between ecosystem CO 2 assimilation and CO 2 losses to the atmosphere. Ecosystem respiration (R eco ), the efflux of CO 2 from the ecosystem to the atmosphere, includes the soil-to-atmosphere carbon flux (i.e., soil respiration; R soil ) and aboveground plant respiration. Therefore, R soil is a fraction of R eco and theoretically has to be smaller than R eco at daily, seasonal, and annual scales. However, several studies estimating R eco with the eddy covariance technique and measuring R soil within the footprint of the tower have reported higher R soil than R eco at different time scales. Here, we compare four different and contrasting ecosystems (from forest to grasslands, and from boreal to semiarid) to test if measurements of R eco are consistently higher than R soil . In general, both fluxes showed similar temporal patterns, but R eco was not consistently higher than R soil from daily to annual scales across sites. We identified several issues that apply for measuring NEE and measuring/upscaling R soil that could result in an underestimation of R eco and/or an overestimation of R soil . These issues are discussed based on (a) nighttime measurements of NEE, (b) R soil measurements, and (c) the interpretation of the functional relationships of these fluxes with temperature (i.e., Q 10 ). We highlight that there is still a need for better integration of R soil with eddy covariance measurements to address challenges related to the spatial and temporal variability of R eco and R soil .