We present an hourly time series of the CO2 concentration profile in the top 20 cm of a boreal forest litter layer at a site in northern Manitoba, Canada. The profile data, measured with an automated sampling system during the summer of 1999, show a pronounced daily cycle, with a small surface CO2 gradient and low concentrations during the day and a large surface gradient and high concentrations at night. The CO2 profile measurements allow us to test two current assumptions built into measurements of ecosystem carbon fluxes. The first assumption is that the flux from the surface to the atmosphere can be calculated using the measured CO2gradient and a calculated value of the diffusive transport coefficient. The behaviour of the surface CO2 gradient suggests that one cannot assume diffusive transport across the moss surface at this site when the friction velocity measured at 30 m exceeds 0.4 m s−1. This condition, associated with turbulent mixing generated by wind shear and/or solar heating of the surface, was often encountered during the day at this site, though rarely at night. During the day, friction velocity and wind speed measured at 30 m height are linearly related, with friction velocity exceeding 0.4 m s−1 when wind speed exceeds about 2 m s−1. At night, wind at the top of the canopy may be laminar, so that the wind speed must exceed 4 m s−1 to cause enough turbulence to raise friction velocity above the 0.4 m s−1 threshold. The second assumption is that changes in soil pore-space CO2 storage can be neglected when correcting eddy covariance measurements for ecosystem respiration that is stored in the ecosystem rather than being mixed into the overlying atmosphere. Our results show that the soil pore-space CO2 profile is not in steady state at the site, but that the magnitude of the corresponding storage flux is small relative to the below-canopy CO2 storage flux. The soil pore-space CO2 storage flux ranges between ±0.4 μmol m−2 s−1, while the below-canopy storage flux ranges between ±20 μmol m−2 s−1. However, the soil pore-space storage flux could be significant relative to the CO2 respiration flux across the soil surface, which we estimate to be in the range of 1–4 μmol m−2 s−1.