J1.2
Turbulent carbon exchange in very stable conditions
Otávio C. Acevedo, Universidade Federal de Santa Maria, Santa Maria, RS, Brazil; and O. Moraes, D. R. Fitzjarrald, R. K. Sakai, M. Czikowski, L. Mahrt, and L. E. Medeiros
Recent studies have shown that turbulent fluxes in very stable conditions can be determined by the use of the multiresolution decomposition. The turbulent fluxes in very stable conditions are very small in magnitude, and have a very short time scale, but are well behaved with proper calculation of the turbulence statistics. A cospectral gap has been shown to exist, clearly distinguishing the turbulent exchange from the larger-scale, more erratic mesoscale fluxes.
We consider data from a deforested site in Amazonia. Significant radiative cooling occur at the open deforested sites, leading to very stable temperature stratification. The nocturnal surface layer at this site can be classified as strongly stable during more than 90 % of the time. Reduced nocturnal mixing often makes the commonly used eddy covariance technique unviable for estimating respiration rates. In the present study, emphasis is given to the analysis of CO2 fluxes, with the purpose of understanding respiration rates in very stable conditions.
Data from the few cases when there is enough turbulent activity at the site show that a good fraction of the carbon fluxes occur at scales larger than the cospectral gap, and are difficult to accurately estimate. Furthermore, and in contrast to what happens for energy and momentum, the carbon fluxes are shown to shift to larger scales as the stability increases. We hypothesize that this is a consequence of the larger time scale for CO2 vertical gradient recovery following mixing events, compared to the same scale for temperature. Temporal evolutions of the vertical gradients are, therefore examined for both mean and turbulent quantities.
Joint Session 1, Stable Boundary Layers 1 (Joint Between 17BLT and 27AgForest)
Monday, 22 May 2006, 8:30 AM-12:00 PM, Kon Tiki Ballroom
Previous paper Next paper