Effects of land-use change on matter and energy exchange between ecosystems in the rain forest margin and the atmosphere

Greenhouse gas and energy fluxes between the land surface and the atmosphere are important aspects for the evaluation of land-use options in tropical areas. Changes in vegetation cover alter the capacity to absorb carbon dioxide and solar radiation from the atmosphere and influence the magnitudes of latent and sensible heat flows to the atmosphere. If happening at a larger spatial scale, land-use change can lead to significant local feedbacks like drought, flooding, soil erosion or shifts in local climate. Up to now only little was known about how typical ecosystems of the rain forest margin areas interact with the atmosphere. We present here results from a sub-program of STORMA (Stability of rain forest margins in Indonesia, SFB-552-B1), to describe the energy, CO2 and water fluxes between a larger tropical area and the atmosphere. Field measurements with the eddy correlation method showed that a Cacao agroforestry system (AFS) was a small source Of CO2, whereas a tropical rain forest was a strong CO2 sink during the one year observation period. In addition, the rain forest evaporated much more water compared to the Cacao AFS. We applied a series of models to extrapolate the results from the field investigations to the region, including so-called SVAT models of differing complexity, and a simple PAR efficiency model to predict net primary production from regional meteorological and remote sensing data. In all cases, we discovered a large spatial variability in CO2 and water vapour fluxes. We investigated the sensitivity of the models to changes in climatic drivers or land-use parameters from the current to a more intensified agriculture. The selected land-use scenarios reduced carbon dioxide sequestration and total evaporation and increased sensible heat fluxes and thus surface temperatures. We conclude that there is still a lack of field observations to better understand and simulate the behaviour of tropical land-use systems. If the necessary detail on model parameter values, which characterise the differences between the land-use systems, is known, the existing model approaches are suitable to investigate consequences of land-use change to regional biogeochemical cycling, water utilisation and climate change.