Improving the Representation of Dual-Phase Diffusion for Land Biogeochemical Modeling in Earth System Models
Mechanistically modeling the biogeochemical processes in land models requires one to resolve the vertical diffusion of different dissolvable volatile gases, such as CH4, O2, and CO2. However, many existing models use ad hoc techniques to represent the dual-phase diffusion of these gases, representing one important structural uncertainty for earth system models. We proposed three equivalent formulations of this dual-diffusion problem and contended that the gas-phase primary form is most straightforward to implement for land biogeochemical modeling. We then solved the gas-phase primary form using a finite volume based algorithm and evaluated its accuracy against three analytical solutions: steady-state CO2 diffusion in soil, steady-state CH4 diffusion in soil, and transient tracer diffusion in two different sequentially aligned medias. The results indicated that our method achieved good numerical accuracy with maximum relative errors less than a few percent. We expect that our method will help land biogeochemical modelers to develop simple and robust algorithms for modeling volatile gas diffusion in variably saturated soil.