Measuring H2O and CO2 fluxes at field scales with scintillometry: Part I – Introduction and validation of four methods

Abstract

This study introduces four methods for determining turbulent water vapour and carbon dioxide flux densities, the evapotranspiration and CO2 flux respectively. These methods combine scintillometer measurements with point-sampling measurements of scalar quantities and consequently have a faster statistical convergence than the eddy-covariance method. The scintillometer measures the friction velocity and stability averaged over space, allowing the time averaging to be a minute or less in homogenous conditions. This paper aims to thoroughly test the methods by analysing their sensitivity to the variables that go into the method and validate the methods with 30-min eddy-covariance data. Introduced are: the Bowen-variance method, the flux-variance method, the structure-parameter method, and the energy-balance method. Sensitivity analysis shows that each method is sensitive to the turbulence measurements of the scalar quantities that are specific to the method, as well as to the friction velocity. This demonstrates that the accuracy of the flux results from a correct representation of the turbulence variables used by the methods. Furthermore, a 30-min flux validation shows that the methods compare well to the independent eddy-covariance fluxes. We found that the structure-parameter method performs best – a low scatter (the correlation coefficient, r=0.99) and a 5% underestimation were observed. Also the other methods perform well, although the energy-balance did not close, because storage terms and CO2 flux were neglected. Furthermore, during the night the variance methods were influenced by non-stationarity in the measurement signal. Finally, we suggest using the correlation coefficients between temperature and scalar quantities to acquire the sign of the fluxes. Data for this study were gathered in May–June 2009 over a wheat field near Merken, Germany, in the framework of the TransRegio32 program.