The water resource is the valuable natural resource for human survival in the world. Surface water and groundwater are the primary constitutions of water resources on earth. Due to intensive anthropogenic activities including industrial, construction, agricultural, and sewage processes, the quality of surface water and groundwater is deteriorated (Paladino et al., 2018; Su et al., 2018). Accordingly, the quality of surface water and groundwater has been the focus of attention for water resource management (Li, 2020; Nasim et al., 2020).

So far, the hydrochemical monitoring of surface water and groundwater has been carried out for improving groundwater management in the world (Yin et al., 2021). Plenty of hydrochemical data provide a robust support for water quality evaluation in the spatial and temporal scale. Various water quality indices (e.g., the WQI and EWQI) have been used to evaluate the quality of surface water and groundwater (Amiri et al., 2014; Lapworth et al., 2017; Zhai et al., 2017). Hence, the spatial and temporal characteristics of the quality of surface water and groundwater have been clarified properly. However, the interaction between surface water and groundwater affecting water quality has remained unclear, failing to explain the dynamic evolution in the spatial and temporal scale.

The interaction between surface water and groundwater can be analyzed using the approaches of hydrochemistry, isotopy, and numerical simulation (Raiber et al., 2019; Marti et al., 2023; Wang et al., 2023). Hydrochemistry and D-O stable isotopes were used to reveal the hydrological relationship between surface water and groundwater in the Notwane River Catchment, Southeast Botswana (Modie et al., 2022). Multiple isotopic tracers (δD-H₂O, δ¹⁸O–H₂O, δ¹⁵N–NO₃ ⁻, and δ¹⁸O–NO₃ ⁻) indicated that shallow groundwater is contaminated with organic fertilizers and subsequently transferred to surface water (Le et al., 2023). The SWAT + gwflow model was constructed to investigate the long-term groundwater–surface water interactions in the Scheldt Basin (Liberoff and Poca, 2023).

Substantial amounts of population are living in basin areas in the world due to the advantages of warm climate and gentle terrain. The Sichuan Basin in southwestern China is the typically large-scale basin within more than 100 million people. The quality of surface water and groundwater has been investigated for decades. However, previous studies mostly focused on the hydrochemical and isotopic information of surface water and groundwater in the basin area (Zhang et al., 2020; Zhang et al., 2021a; Zhang et al., 2021b). Seldom studies investigate the dynamic interaction and transport processes between surface water and groundwater at the boundaries of plains and mountains. It is known that the contamination would be transferred between surface water and groundwater (Xu et al., 2022). Among those pollutions, the contamination produced by the mining projects and tunnel constructions in the mountainous area is a serious environmental problem (Tomiyama and Igarashi, 2022). The water system is interrupted by excavation, and the unpredicted groundwater flow may pollute the fresh groundwater in mining sites and surface water system in the plain (Wu et al., 2018), especially the transport of heavy metal in the wastewater (Santana et al., 2020). The recent investigation pointed out that mountainous water system in Sichuan Province is heavily contaminated by the regional mining and tunneling industry (Sun et al., 2023). Hence, a hydrological model of large-scale catchment from the basin margin to basin center is built in this study. Afterward, the dynamic interaction between surface water and groundwater is simulated under different conditions of climate precipitation and terrain slope. Finally, the contamination in surface water and groundwater can be evaluated under different conditions. The achievements of this study would significantly contribute to evaluating the quality of surface water and groundwater in the catchment of the Sichuan Basin.

To investigate the issue of groundwater contamination from mountains to lowland rivers, a 2D numerical model is constructed in this study. The groundwater flow is steady-state since our model is synthetic to represent the general situation. Flow is simulated using MODFLOW (Harbaugh, 2005) with the Darcy equation (Eq. 1), and the solute transport is simulated in MT3DMS (Zheng and Wang, 1999) to depict advection and dispersion (Eq. 2). ∂ ∂ x K ∂ H ∂ x + ∂ ∂ z K ∂ H ∂ z + w = 0 , (1) where K is the value of hydraulic conductivity [LT⁻¹]; H is the hydraulic head [L]; and w is the volumetric flux per unit volume representing sources and/or sinks of water [T⁻¹]. ∂ ∂ x i θ D i j ∂ C ∂ x j − ∂ ∂ x i θ v i C + q s C s = ∂ θ C ∂ t 2 , (2) where θ is the porosity of porous media [dimensionless]; v _i is the seepage velocity [LT⁻¹]; C is the solute concentration [dimensionless]; C _s is the solute concentration of water entering from sources or flowing out from sinks [dimensionless]; D _ij is the hydrodynamic dispersion coefficient tensor [L²T⁻¹]; and q _s is the volumetric flow rate per unit volume of aquifer representing the fluid source (positive) and sink (negative) [T⁻¹].

This study aims to investigate the effect of slope and precipitation on the transport of contaminants from the mountainous area to lowland river. Figure 3 illustrates examples of simulation results for steady-state groundwater levels. At a precipitation level of 1200 mm/yr, the highest groundwater head level is approximately 300 m for the 40° slope angle, 320 m for the 30° slope angle, and 340 m for the 20° slope angle (Figures 2A–C). It indicates that the groundwater head level increases with a steeper slope angle. On the other hand, the precipitation also controls the groundwater level, in which the groundwater head increases with precipitation from 600 mm/yr to 1,000 mm/yr (Figures 2D–F). Hence, it can be inferred that groundwater head levels possessed a positive relationship with precipitation while a negative relationship with the slope angle.

The process of solute transport was simulated from the mountainous area to the river based on the steady-state flow field (Figure 3). Figure 4 presents the arrival time and total contaminant mass to the river for each simulation case. The arrival time analysis reveals that pollutants take approximately 70 years to reach the river. Specifically, the case of 20° slope angle at a 1,200 mm/yr rainfall recharge rate exhibits the earliest arrival time at ∼65 years. In contrast, the case of 35° and 40° slope angles at a 600 mm/yr rainfall recharge rate show the slowest solute transport with the arrival time at ∼75 years. Arrival time decreases with increasing precipitation, indicating that higher rainfall recharge rates promote the speed of solute transport. Furthermore, it increases with slope angles, highlighting that the gentler slopes are more conducive to contaminant transport. On the other hand, the total mass of pollutants discharged into the river is greatly influenced by the precipitation. The mass of contaminants discharged into the river increases from 45 to 67 with the increasing rainfall recharge rate over a 20-year period. However, the relationship between total mass and slope angle is less apparent, suggesting that slope angles have a limited impact on the mass of contaminants transported into the river.

The findings of this study indicate the importance of evaluating contaminant transport between the mountainous area and basin plain, suggesting that a comprehensive analysis with multiple factors are needed in studying the contamination interaction between groundwater and surface water. In addition, the protection of groundwater resources in the mining sites and tunnels should be stressed as the contaminants possibly transport with groundwater from mountains to the surface water in the basin. Engineers should pay more attention to the groundwater flow system and water–rock interaction around the mining sites and tunnels, as well as techniques of water drainage and pollutant deposition in the underground engineering.

This study conducts simulations to analyze contaminant transport from the mountainous area to the lowland river. The investigation considers multiple cases with varying precipitation and slope angles along the gradient from the mountain to the plain, with a focus on calculating the total mass and arrival time of contamination. The simulation results reveal that both groundwater level and solute transport are influenced by precipitation and slope angles. The main findings are listed below:

(1) The general groundwater head demonstrates an increase with higher precipitation and a decrease with steeper slope angles, ranging from 300–340 m.