Comparative study on two drying-rewetting algorithms of groundwater model cells

When simulating drying-rewetting process of grid cells in numerical groundwater modeling using the block-centered finite-difference approach, the models is highly probable to run into non-convergence, which could greatly affect the applicability of groundwater models. This study used ideal case and practical simulation in Denmark to comprehensively compare the simulation capabilities and characteristics of two algorithms, namely the empirical trial (ET) method proposed by MODFLOW and the always active cell (AAC) method proposed by COMUS, in the drying-rewetting simulation. The results show that: (1) For the ET method, the selection of parameter combination has a significant influence on the model convergence and the simulation results. It is compulsory to continuously optimize the parameter to avoid model failures such as non-convergence and large simulation errors when using the ET method, which greatly increases the difficulty of groundwater model application and time cost. (2) The simulation results from the AAC method are more reliable than those from the ET method. Theoretically, the effect of the AAC method is equivalent to the optimal parameter set in the ET method. Therefore, parameter optimization is no longer needed in the AAC method, which can effectively reduce the difficulty of using groundwater models and meanwhile reduce the uncertainty of simulation results. (3) The numerical accuracy of the intercell horizontal hydraulic conductance in the AAC method is consistent with that of the classical harmonic average method, demonstrating that the AAC method can also be used in the simulation without the drying-rewetting process. In summary, the AAC method is more suitable for simulating the drying-rewetting process of groundwater model cells and is expected to be more widely used in groundwater numerical simulation.