Wetland methane fluxes : upscaling from kinetics via a single root and a soil layer to the plot

The aim of this thesis was to increase the understanding of plot scale relations between CH 4 fluxes and environmental variables in wetlands. Theories of microbial and chemical conversions were taken as starting point, as a literature review showed that it is hard to relate methane production and oxidation directly to environmental variables. These theories only apply under homogeneous conditions at the kinetic scale (here about 1 mm) and were linked to plot scale CH 4 fluxes by stepwise scaling up.At the kinetic scale a CH 4 production model was developed, comprising anaerobic C-mineralisation, electron acceptor reduction, methanogenesis and methanogenic growth, of which the last process is probably not important in wetland soil. Application of this model to anaerobic incubation experiments with peat soil suggested that organic peat may act as terminal electron acceptor, using a substantial amount of anaerobically mineralised C. At the single root scale CH 4 dynamics were explained with coupled reaction-diffusion equations for CH 4 , oxygen (O 2 ), molecular nitrogen, carbon dioxide and an electron acceptor in oxidised and reduced form. Included conversions were: aerobic respiration, C-mineralisation, CH 4 production and oxidation, electron acceptor reduction and re-oxidation. Root gas transport was described with first order gas exchange over the root surface. Bubble formation was modelled with simultaneous liquid-gas equilibria of all gases and bubble export with a descriptive relation with bubble volume. The model was simplified by assuming quasi steady-state for O 2 and by spatially averaging the other compounds. These simplifications had little effect on simulated CH 4 dynamics and therefore the simplified model was used at the next higher level.At the soil layer scale the CH 4 dynamics were calculated with a weighed set of single root systems with different distances to the next root. These weights were calculated from the root architecture, conserving the probability density function of the distance to the nearest root. The model was simplified by averaging over the single root systems. This had some effect onCH 4 production and CH 4 transport, but little on CH 4 emissions.At the plot scale, temporal water unsaturation was accounted for with Richards' equation. The soil layer models were extended to the plot scale by incorporating vertical transport of the compounds by diffusion and mass flow. SimulatedCH 4 fluxes were of the same order of magnitude as measured fluxes. They were sensitive to several uncertain parameters, indicating that predictive process modelling of CH 4 fluxes is not possible yet. Heterogeneities within a soil layer seem to be less important than heterogeneities between soil layers. This may be explained by a weaker effect on the OO 2 input into the soil.CH 4 fluxes result from the electron donor input minus the electron acceptor input and changes in storage of electron donors, electron acceptors and CH 4 in the soil. The developed models showed that the changes in storage are the result of a number of uncertain processes. Hence, the most stable relationships between CH 4 fluxes and environmental variables may exist at larger time scales.To conclude, a coherent set of models was developed that explicitly relates processes at the kinetic, single root and soil layer scale to methane fluxes at the plot scale.