Numerical modeling of ammonia nitrogen transport in laboratory column experiments

Ammonia nitrogen is very harmful and has been detected in groundwater globally. To describe the migration of ammonia nitrogen in the vadose zone accurately, we herein compose and analyze breakthrough curves (BTCs) of ammonia nitrogen for experiments and models. The main experiments of adsorption and column were presented for the process of ammonia nitrogen. Also, the typical experimental media of silty loam and silty clay were taken from Changchun in China. Of three models (linear, Freundlich and Langmuir), the Freundlish can be fitted best to the adsorption isotherms of both soils. Based on fitted two parameters of Freundlish model, the maximum adsorption capacity of 833.33 μg/g for silty loam and 1428.57 μg/g for silty clay was calculated. BTCs for ammonia nitrogen show obvious non-Fickian dispersion comparing with its for chloride. BTCs for chloride and ammonia nitrogen are derived and transport parameters are estimated using three approaches: the equilibrium advection-dispersion equation (ADE), two region model (TRM), and continuous time random walk (CTRW). It was found that all the models fit the transport process for chloride well, but the CTRW is better. While the concentration breakthrough curve of ammonia nitrogen is described effectively by CTRW. Furthermore, for ammonia nitrogen as a reactive solution, the CTRW with sorption term can capture the full evaluation of the tracer-breakthrough curve and late-time tailing during leaching, while the traditional ADE and TRM cannot. In particular, the physical heterogeneity and chemical reaction are demonstrated to account for its non-Fickian transport. Finally, through inverse calculation of the maximum adsorption capacities of both media, the interception capacities were found to be far greater than their adsorption capacities. Our results reveal the anomalous concentration variance and shows a prediction model of ammonia nitrogen, helping decision-maker assess and manage the groundwater pollution.

This study was supported by the 111project (No. B16020), the National Natural Science Foundation.