Author(s): Zhibo Zhang; Zhi Li

Linked Author(s):

Keywords: Subsurface contaminant simulation; Fully integrated numerical model; Model verification

Abstract: The unsaturated zone serves as a critical nexus for the exchange of water and energy between atmospheric water, surface water, and groundwater, acting as the source of sustenance for terrestrial plants and a conduit for surface pollutants to infiltrate groundwater. It holds significant importance within the entire hydrological cycle system. The movement of soil moisture is extremely complex, making it crucial to accurately describe the laws of water and solute transport in saturated and unsaturated zones, and to efficiently assess the impact of soil water and solutes on the groundwater environment. Simulations of soil water and solute transport in the unsaturated zone primarily rely on numerical solutions of the Richards equation and the convection-dispersion equation, as well as various simplified equations. However, the strong nonlinearity of the Richards equation imposes strict requirements on the spatial and temporal discretization resolutions of the numerical schemes, leading to high computational costs. Due to the differences in movement characteristics between saturated and unsaturated zones, there are relatively few models that adopt a fully coupled approach. High-performance computing (HPC) technology, as a powerful tool, has become an effective means of enhancing computational efficiency. This study, based on the existing high-performance groundwater dynamics model SERGHEI-RE, develops a fully integrated model for solute transport in both saturated and unsaturated groundwater using the Kokkos heterogeneous programming framework, aiming to achieve efficient numerical simulation of the entire process of groundwater movement and solute transport. By combining classic examples from literature, the results were validated against existing groundwater solute transport models HYDRUS, showing that the SERGHEI-RE-RT can reasonably calculate water flow and solute transport processes even under complex boundary and variable soil conditions. Furthermore, simulation speed tests of the coupled model under conditions such as single-core, multi-core CPUs, and single-GPU indicate that the model can efficiently simulate water flow and solute migration with relatively low computational costs.

DOI: https://doi.org/10.64697/978-90-835589-7-4_41WC-P2032-cd

Year: 2025