Net carbon uptake and coupling from a tropical wet forest

Eddy covariance (EC) measurements of CO2 were conducted in tropical lowland wet forest at the La Selva Biological Station in Costa Rica (10.26 N, 83.59 W). We also examined the wind statistics and its implications for coupling of the canopy to the atmosphere. The EC system was based on a closed path infrared gas analyzer and sonic anemometer mounted on a 42 m tower. Prevailing winds are easterly from the Caribbean with little anthropogenic influence. The tower footprint is primary upland forest. We estimated the annual net ecosystem exchange based on EC and quarterly relationships of respiration and uptake rates. EC data was filtered for U*> 0.15 m s^1 and no rain events. Ecosystem-level respiration was found to have a Q10=5.1. During 1999, we estimated the forest was a sink for carbon, 189 g C m^-2 y^-1. Approximately 2/3 of this carbon was sequestered during the dry season. The results suggest that contrasting climatic factors control carbon uptake during the different seasons.

We examined the sub-inertial range with spectra analysis and a typical -2/3 slope indicates a dissipation of eddys (thus a coupling of the canopy to the atmosphere). We found this relationship during the day under both convective and mechanical turbulent conditions with U* > 0.12 m s^-1^. However, during nighttime, under thermally stratified boundary-layer with U* £ 0.12 m s^-1^ we found a similar -2/3 relation indicating longer waveforms interacting with the canopy such that mass and energy are still being exchanged likely due to differential heating and cooling of the canopy and changes in roughness length with direction. Residuals from the energy balance (EC derived estimates v. net radiometry) were not a function of U*. Moreover, we found departure in daytime measures of the stability function from the Dyer similarity function, suggesting that long aerodynamic roughness lengths @ 2.0 m and at times buoyancy terms become dominate. Direct implications of this study include: (1) the need to examine localized similarity functions and the function of shear to buoyancy terms for closure in the energy balance (2) using spectral data to correct for signal frequency loss may not be appropriate for complex canopies.