| Summary: | This study evaluates the impact of post-field data processing methods on eddy covariance flux estimates and the resulting energy balance residual. To that end, a dataset from the LITFASS-2003 field campaign was analysed using an experimental software package. Widely discussed issues in data processing like an adequate flux averaging time, coordinate transformations and alternative approaches for the correction of density effects were examined. The impact of all the single processing steps on the turbulent flux estimates of sensible heat, latent heat and CO2 as well as the impact on the resulting energy balance residual is demonstrated. The mean energy balance residual could be reduced by 17 % compared to 30-min covariances without any conversions or corrections if all necessary procedures were applied to the test dataset. Important procedures for the reduction of the experimental energy balance closure problem were the correction for spectral loss and the correction for density effects. Furthermore, the energy balance residual vanished almost completely if the covariance averaging time was extended from 30 minutes over 24 hours to five days. The low-frequency flux contributions could be explained through effects which were caused by the strong heterogeneity of the landscape surrounding the measurement site. The dependence of CO2 flux estimates on the post-field data processing was stronger. Their mean value was changed by a factor of 2 through the correction for density effects.
|
|---|