On turbulent exchange processes over Amazonian forest / Celso von Randow.

Tese (PhD.)- Wageningen University, 2007

This thesis deals with turbulent exchange processes of heat, humidity and carbon dioxide over Amazonian forests. The atmospheric boundary layer over Amazonia frequently contains slowly-moving large eddies induced by strong convective motions or local circulations related to the heterogeneity of the surface. To better understand their influence, it is studied how turbulence statistics depend on different time scale and spatial scale classes, decomposing the turbulence signals using multi-resolution (wavelets). Largest contributions to measured fluxes occur in turbulent scales (structures with length scales up to 1000 m and time scales up to 15 min). Low-frequency motions (larger eddies and mesoscale motions), however, can contribute with up to 30 % to the total exchange under weak wind conditions. The influence of low-frequency motions on similarity and correlations between turbulent signals and the implications for the application of Monin-Obukhov similarity is investigated. For the estimation of heat fluxes by the flux-variance method it is found that reasonable results only occur when the correlation coefficient between vertical wind and temperature (rwT) is above 0.5. The latter quantity decreases when the influence of low-frequency motions in the surface layer is high, and in these cases the surface layer is different from the textbook descriptions. Additionally, the use of Large Aperture Scintillometry (LAS) over Amazonia is explored, in comparison with the eddy covariance (EC) method. As the LAS provides a measurement that represents a weighted spatial average of the turbulent eddies along the path, it is not necessary to use a long time period to sample a number of eddies, and the averaging time scale can be reduced. The results show that the EC fluxes are often lower than the LAS. The difference increases with increasing non-stationarity conditions and decreasing correlation coefficient In general, the results suggest that one single eddy covariance system is not able to capture quasi-steady large-eddies that significantly contribute to surface-atmosphere exchange over Amazonian forest. Apart from possible horizontal flux divergence at heterogeneous terrain, these factors may explain the failure to close the surface energy balance in complex terrain.