Pore network modeling as a tool for determining gas diffusivity in peat

Peatlands are significant modulators of biogeochemical cycles and important carbon stocks on a global scale, and they may become large sources of the greenhouse gases (GHG) carbon dioxide and methane because of their vulnerability to management practices and changes in climate. Because gas exchange between peat and the atmosphere occurs primarily via diffusion, a proper knowledge of the gas diffusion rate is essential for correct estimation of the amount of GHG emissions from peatlands. Diffusion is controlled by the structure and connectivity of peat pore space. Pore network modeling (PNM) is an efficient method for the pore-scale description and simulation of transport processes in porous matter and, therefore, a useful tool for the assessment of gas diffusivity in peat, as it explicitly illustrates the relationship between the peat microstructure and the gas transport properties on a macroscopic scale. PNM can also be used to simulate time-dependent soil biogeochemical processes, such as GHG production and consumption.
We extracted interconnecting macropore (diameter greater than 0.1 mm) networks from three-dimensional X-ray micro-computed tomography (µCT) images of peat samples from three depths and simulated steady-state diffusion in the networks using PNM. We then compared the obtained soil gas diffusion coefficients to those determined experimentally from the same samples. The gas diffusivity measurements were made using the diffusion chamber method under different water contents adjusted in a pressure plate extractor. 
The measured soil gas diffusivity was lower in deeper layers because of decreased air-filled porosity and pore connectivity. Nevertheless, the diffusion rates were not extremely low close to saturation, which may imply that connected air-filled pathways for gas diffusion are present in peat even in wet conditions. The pore network simulations were able to reproduce the experimentally determined gas diffusion dynamics rather well. This also implies that the topology and the dimensions of the pore space of most of the peat samples were adequately represented by the network objects. Therefore, the combination of µCT and PNM can be considered a potential alternative to the assessment of soil gas diffusivity through traditional laboratory measurements. However, further research is needed on gas diffusivity in different peat types over a wide water content range. Furthermore, the presented approach provides a basis for mechanistic simulation of GHG cycling in soils.