Numerical modelling for evaluation of the nitrogen removal rate in hyporheic zone

Abstract

The hyporheic zone (HZ), the region beneath or alongside a streambed where active groundwater and surface water mix, plays a vital role in the stream ecosystem. Reactions in the HZ such as denitrification and nitrification have been examined in previous studies. However, those numerical models are lack of consideration for the reaction zones for aerobic and anaerobic changes due to the reactions consuming the dissolved oxygen (DO) in hyporheic flux. In order to simulate nitrogen concentration changes in the HZ more accurately, this study proposes a method of evaluating the nitrogen removal rate in the HZ through numerical modeling. Firstly, a basic two-dimensional numerical model following previous simulation models in the HZ, which only couple flow conditions with biochemical reactions, is proposed to consider both nitrification and denitrification, but ignoring the changes generated by the reactions. Next, the zones for different reactions are determined in an improved model under the assumption that related environmental variables (i.e., the DO) will be considered to delineate the boundary between nitrification and denitrification zones and to identify a transition zone where either reaction might take place. The changes of reaction zones through the whole process can be controlled by the characteristic variable of hyporheic flux, and this variable can be selected differently for different reaction processes. In this study, the characteristic variable is determined as median residence time. Using this model, the accuracy of the nitrogen simulation in the HZ can be improved. To overcome the shortcoming that more information about biochemical reactions in the HZ is required to use the improved model, a new model that couples the basic model and genetic programming (GP) is proposed to optimize the simulation results of the basic model and allow for real-time forecasting. The results show that the improved model performs better than the basic model, but the model coupling the basic model with GP performs best. In addition, the function of the HZ in nitrogen removal is examined through a case study of four scenarios, leading to the conclusion that the HZ has a higher nitrogen removal rate when water quality is neither too poor nor too good. Therefore, even though the HZ facilitates nitrogen removal, sewage should still be treated to a certain level before being discharged into rivers. Overall, this study enhances our understanding of the HZ, and can benefit the restoration and management of HZs and streams in the face of the continual degradation caused by human activity.