Measuring H2O and CO2 fluxes at field scales with scintillometry: Part II – Validation and application of 1-min flux estimates.

Abstract: This paper evaluates four methods to obtain accurate averaged flux estimates under conditions of non-stationary turbulence. In Part I (Van Kesteren et al., 2012), we introduced and evaluated these four combined methods for 30-min averaging intervals, notably the flux-variance method, the Bowen-variance method, the structure-parameter method, and the energy-balance method. The aim of this paper, Part II, is to validate the accuracy of the 1-min flux estimates of the CO2 flux, FCO 2, and the evapotranspiration/latent-heat flux, L v E. Furthermore, we use the 1-min fluxes to investigate flux and vegetation responses under conditions of non-stationary turbulence. Using several validation methods, we show that both the eddy-covariance method and the energy-balance method are unsuitable for estimating fluxes over 1-min averaging intervals. The three other combined methods are more successful in determining 1-min fluxes. The random error is approximately half that of the eddy-covariance method, but still some issues limit the success. The Bowen-variance method has a +0.09 systematic error and moreover, 30% of the data had to be omitted, because the method requires more stringent conditions. Furthermore, the flux-variance method has a −0.15 systematic error. The structure-parameter method performs best of all methods and accurately resolves 1-min fluxes. With this method, we do a final validation with a different data set and show that also under dry conditions the method accurately resolves FCO 2, although L v E was more difficult to resolve. In the last part, the structure-parameter method is successfully applied under conditions of non-stationary turbulence. We show that L v E and FCO 2 have a different step response upon abrupt changes in solar radiation, because different processes drive these fluxes. Also, we observe a 2-min time lag between solar radiation and 1-min fluxes and show the relevance of taking this into account for determining light–response curves of the plants for both 1-min and 30-min averaging intervals. Furthermore, we show the relevance of 1-min fluxes for studying the light–response curves of plants for conditions with different temperature and humidity. Finally, we show that accurate estimates of 1-min averaged canopy resistances can be determined via the resistance expressions for sensible heat and L v E. As such, we show that vegetation can indeed modify its canopy resistance significantly within several minutes. [Copyright &y& Elsevier]