Measuring water-vapour and carbon-dioxide fluxes at field scales with scintillometry

Scintillometry is a measurement technique that has proven itself to be of great value for measuring spatial-averaged fluxes of sensible heat, momentum, and evapotranspiration. Furthermore, for crop fields (field scales), scintillometry has been shown to accurately determine the sensible-heat and momentum flux over time intervals as short as 6 seconds. As a consequence, interests in scintillometry are growing and scintillometers that determine sensible-heat fluxes and momentum fluxes have become commercially available. This thesis deals with two aspects of scintillometry. First, after a general introduction of scintillometry, measurement errors that have been observed in the large-aperture scintillometer from Kipp&Zonen and in the SLS field-scale scintillometer from Scintec are evaluated. For both scintillometer types, we discuss the variability in the measurement errors among different instruments and, where possible, we give solutions to remove these errors. Furthermore, we present the results of a prototype scintillometer that was developed as part of the research project. With our proposed design, we aim to overcome the measurement errors in the Scintec scintillometer and extend the applicability of the field-scale scintillometer to paths that are longer than 200 m. Second, we extend the application of field-scale scintillometry to the flux measurements of latent-heat, carbon-dioxide, and other passive scalars. Until now, scintillometers could not be used for measuring passive-scalar fluxes over crop fields and we show that with our extended methodology these fluxes can be accurately determined over time intervals as short as 1 minute. The methodology is based on a combination of scintillometer measurements and additional high-frequency scalar measurements and works under conditions of homogeneous turbulence, i.e. single crop fields. We introduce four methods, notably the energy-balance method, the Bowen-variance method, the flux-variance method, and the structure-parameter method. Using several validation methods, we show that the energy-balance method is unsuitable for estimating scalar fluxes over 1-min averaging intervals. The Bowen-variance and flux-variance method perform better and the structure-parameter method accurately resolves 1-minute fluxes. Thus, with this methodology fluxes can be resolved with a high temporal resolution, making it possible to study vegetation in a natural environment under non-stationary conditions. This allows us to show that the wheat vegetation affects fluxes upon changes in solar radiation in time periods clearly shorter than 30 minutes and that the canopy resistance can change significantly within several minutes.