Different Eddy Sampling methods derived from the Eddy Covariance method are used to measure trace gas fluxes in the boundary layer when fast sensors are not available. Sampling in Relaxed Eddy Accumulation (REA), Hyperbolic Relaxed Eddy Accumulation (HREA) and Disjunct Eddy Covariance (DEC) was simulated using high resolution datasets from the EBEX-2000 Experiment (Oncley et al., 2000) and data acquired over a high grown spruce stand at Waldstein, Germany (Wichura et al., 2001). These simulations allowed to quantify sources of error in REA, HREA and DEC methods stemming from underlying assumptions and depending on the meteorological conditions.

Methods based on Relaxed Sampling are indirect methods for flux measurements, because they rely on a parameterization in which the so called factor b is determined from a second scalar quantity which can be measured with high temporal resolution, e.g. temperature or water vapor (Businger and Oncley, 1990). In agreement with earlier studies (Oncley et al., 1993) simulation of ideal REA resulted in minimum relative flux errors when a wind deadband of was applied. Simulation of HREA identified periodic large errors from violation of scalar similarity required for the determination of the hyperbolic deadband and the b-factor.

Disjunct Eddy Covariance (DEC) is a direct method for flux measurements (Lenschow et al., 1994; Rinne et al., 2001). Results from simulation of DEC showed no evidence for systematic error. The random error was quantified and found to depend mainly on the number of samples acquired when comparing results for buoyancy flux, humidity flux and CO2 flux.