, Volume 91, Issue 2, pp 259-280

Autoregressive filtering versus linear detrending in estimation of fluxes by the eddy covariance method

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


The application of autoregressive running mean filtering (RMF) and linear detrending (LDT) in the estimation of turbulent fluxes by the eddy covariance method is analysed. The systematic, as well as the random, errors of the fluxes arising from filtering and/or limited observation time effects are described. To observe negligible systematic errors in fluxes, the RMF has to be applied with moderately long time constants. However, the obtained flux values are subject to increased random errors during periods of non-stationarity and the method leads to systematic overestimation of variances. These shortcomings are not inherent in the LDT approach, which is recommended for use. But the systematic errors of fluxes due to LDT are not negligible under certain experimental conditions and have to be accounted for. The corrections are important because the relatively small errors in short-period fluxes can translate to significant errors in long-period averages. The corrections depend on the turbulence time scales, which should be preferably estimated as ensemble mean variables for a particular site.