**1. Introduction**

Contamination of the environment with various trace metals has been of significant concern for many decades due to their potential to cause deleterious impacts on exposed wildlife and humans. Trace metals have several unique chemical properties that dictate their environmental fate and bioavailability. Trace metals such as copper, nickel, zinc, and lead are inclined to interact with other elements and organic molecules present in the environment. Assessments of environmental fate and risks of trace metal exposure to wildlife and humans historically only accounted for these interactions superficially and are most commonly based on total metal concentrations. Thus, there is a need to develop methods and strategies to overcome this limitation and help improve risk assessments for both anthropogenic metal contamination, e.g., with copper, and geogenic background contamination, e.g., with nickel.

Copper reacts strongly with various functional groups present in soils and sediments, such as iron oxides and manganese oxides [1]. Localized deposits can be caused by

**Citation:** Prajapati, S.; Sabokruhie, P.; Brinkmann, M.; Lindenschmidt, K.-E. Modelling Transport and Fate of Copper and Nickel across the South Saskatchewan River Using WASP—TOXI. *Water* **2023**, *15*, 265. https://doi.org/10.3390/w15020265

Academic Editors: Bommanna Krishnappan and Andrea G. Capodaglio

Received: 30 November 2022 Revised: 3 January 2023 Accepted: 5 January 2023 Published: 8 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

anthropogenic activities such as mining, managing municipal and industrial waste, using copper as a pesticide in agriculture, and water treatment [2]. As a result, copper may enter the freshwater system naturally, as a result of human activity, or due to the corrosion of pipes and fittings in water distribution systems [3,4]. Copper is utilised in a wide variety of products, including pipes, paints, refinery oils, and construction materials, because of its malleability, conductivity, alloying potential, and resilience to corrosion and wear [2,4,5]. Copper exists in four different oxidation states: Cu (elemental copper), Cu (I) (cuprous ion), Cu (II) (cupric ion), and Cu (III) [6]. In water and sediment systems, Cu (II) is a species that is frequently found [7]. Since copper (I) is unstable, it easily oxidises to copper (II), which then undergoes additional redox processes and becomes hydrated in the presence of water molecules. The hydrated Cu (II) ion binds to dissolved organic and inorganic molecules as well as particulate debris [8]. In soil, copper does not often bioaccumulate or even hydrolyze in any form [9]. Cu speciation is complicated and dependent on a number of variables, including water chemistry (dissolved oxygen, pH, and redox potential), water hardness, and sediment–water interaction [10]. The toxicity of copper in freshwater is mainly driven by various chemical parameters and water quality. According to the biotic ligand model (BLM) proposed by [10], metal toxicity is driven by the accumulation of metal at a discrete site of action (biotic ligand). Metal speciation results in formation of inorganic and organic complexes [11]. As discussed above, Cu (II) adsorbs to particulate matter and forms dissolved complexes with both organic and inorganic ligands. This ability of Cu (II) is a major contributor for it to be the major contributor to copper toxicity (ECCC 2021). Water quality monitoring datasets from Environment and Climate Change Canada (ECCC) were compiled for different provinces and territories. Total copper concentrations of Canadian jurisdictions varied with an overall range from 0.002 (Ontario) to 5723 µg/L (Saskatchewan). The median copper concentrations ranged from 0.3 µg/L for Prince Edward Island (PEI) to 1.8 µg/L for Manitoba.

Nickel (Ni) is one of the most common elements on Earth and is widely distributed in the environment [12]. Ni occurs mostly in the mineral form at an average concentration of ~75,000 µg/kg in the Earth's crust. Ni is known to have a widespread distribution in the environment and is an essential component for many industrial and commercial uses. Various uses of nickel include electroplating, as a catalyst for fat hardening, coloring ceramics, electrical components, ammonia adsorption processes, and other metallurgical operations [13]. Nickel is released into the environment in various forms, predominantly from natural sources and few anthropogenic activities. Natural weathering and erosion of geological materials release nickel into surface waters and soils in Canada. Forest fires can be short-term but intense sources [14]. Mining, smelting, petroleum refining, and manufacturing industries are other major emitters [15]. Nickel moves in both particulate and dissolved forms in natural waters. The transport and fate of nickel in freshwater rely on various factors such as the pH, redox potential, ionic strength, type, and adsorption type [16]. Nickel toxicity is dependent upon the route of exposure and the solubility of a nickel compound [17]. The movement of nickel in the environment is high in acidic organic-rich soils, which can lead to groundwater contamination [18]. In sediments and suspended solids, most of the nickel is distributed among organic materials, precipitated and coprecipitated particle coatings, and crystalline particles.

Studying metal speciation in freshwater systems is challenging, as the applied methodologies need to be very sensitive to be able to recognize minute differences in relative proportions of metal species at trace concentration levels. Various models have been used to compare predictions of metal speciation to experimental values, which has been documented in a few studies [19–22]. Models such as the biotic ligand model (BLM), the Windermere humic aqueous model (WHAM), the NICA–Donnan model, and WASP—TOXI [23] have been developed to study or predict metal speciation in the environments at varying levels of complexity. BLMs determine metal speciation and predict metal toxicity to biota in aqueous systems using computational modelling. WHAM is based on the Humic Ion-Binding Model and assumes that proton and metal complexation occurs at two groups

of discrete sites with strong and weak binding affinities [24]. The NICA–Donnan model considers carboxylic- and phenolic-type groups for determining site strengths [25]. In a study [26], modelled in situ concentrations of Cd, Cu, Ni, and Pb were compared with measured values by employing different speciation techniques, specifically the Humic Ion-Binding Model VI (WHAM 6) and the NICA–Donnan model. Both the models were found to be reasonably accurate and performed consistently for metal speciation in freshwater. However, concentrations of total dissolved Cu and Pb were underestimated by a large magnitude. While this finding is generally promising, the complexity of identifying anthropogenic versus geogenic sources of contamination, impacts of water chemistry/quality on the fate of metals, hydrology, and morphodynamics requires an integrated modelling strategy that accounts for all of these processes.

In this study, we used the Water Quality Analysis Simulation Program (WASP) to model the fate of the metals nickel (predominantly from geogenic background) and copper (predominantly from anthropogenic sources) between water and sediments in the South Saskatchewan River through the sampling periods of 2020 and 2021. With an average discharge of 277 m3/s (min: 68 m3/s, max: 731 m3/s) in 2020 and 134 m3/s (min: 68 m3/s, max: 330 m3/s) in 2021, the hydrology in these two sampling years was noticeably different (both measured at the outlet of Lake Diefenbaker). Additionally, Southern Saskatchewan experienced a unique pattern of precipitation in the two sampling years, with an average of 297.4 mm in 2020 and 180.7 mm in 2021 (ECCC historical weather data). This study makes use of trace metal concentrations and other physicochemical variables recently published by our group [27]. Comparing the predictive power of WASP for trace metal concentrations in sediment and water across two hydrologically different sampling periods, encompassing both a flood and a drought year, bears a considerable potential to test the model across a wide range of hydrologic conditions and, thus, the potential utility of this model in forecasting, e.g., with the goal of testing the impacts of climate change scenarios.

#### **2. Materials and Methods**

#### *2.1. Study Site*

The Saskatchewan River system, which originates in Alberta's Rocky Mountain headwaters and flows through Canada's plains, is made up of the South and North Saskatchewan Rivers. The South Saskatchewan River Basin, which includes the major cities of Saskatoon, Swift Current, Red Deer, Calgary, Lethbridge, and Medicine Hat, is situated in the southern regions of Alberta and Saskatchewan (Figure 1). Big Stick Lake, Bow River, Oldman River, Red Deer River, Seven Persons Creek, South Saskatchewan River, and Swift Current Creek are all parts of the South Saskatchewan River Basin [28]. *Water* **2023**, *15*, x FOR PEER REVIEW 4 of 16

**Figure 1.** Sampling locations in the South Saskatchewan River Basin, Saskatchewan [27].

charge of 280 m3/s at the Saskatchewan River Forks. It covers a watershed of approxi-

in Alberta and Saskatchewan **[**29**]**. Agriculture accounts for two-thirds of the land cover

The change of heavy metals in the South Saskatchewan River system was studied using the Water Quality Analysis Simulation Program 7.52 (WASP) coupled with HEC-RAS. The WASP model could be used to simulate many different aquatic systems. The application could simulate hazardous water contamination using the concepts of mass, momentum, and the conservation of energy. The hydrodynamics of the model domain was simulated using the hydraulic model HEC-RAS 6.3. Since HEC-RAS was developed by the U.S. Army Corp of Engineers (USACE), all users and organizations have access to a free version of the model. HEC-RAS is available to everyone and does not require a license, making it a key benefit of utilizing it as a modelling tool. HEC-RAS can model flow, sediment transport, water quality in one-dimensional (1D) steady and unsteady flow, and two-dimensional (2D) unsteady flow of rivers. The model uses geometric data representation and geometric and hydraulic computer algorithms for a network of river channels. The HEC-RAS program can simulate an input flood using either a (1D) unsteady

Flow data were acquired from available flow data from the Water Survey of Canada (WSC) station at Lake Diefenbaker, which has an operational gauge (records available

The Red Deer, Bow, and Oldman rivers meet at the confluence of the South Saskatchewan River, which is fed by the glaciers of the Rocky Mountains. From its source, the

, of which 1800 km<sup>2</sup> are in Montana, United States, and 144,300 km<sup>2</sup> are

**Figure 1.** Sampling locations in the South Saskatchewan River Basin, Saskatchewan **[**27**]**.

flow model or a (2D) unsteady flow model following these parameters.

mately 146,100 km<sup>2</sup>

*2.2. Modelling Approach*

**[**30**]**.

The Red Deer, Bow, and Oldman rivers meet at the confluence of the South Saskatchewan River, which is fed by the glaciers of the Rocky Mountains. From its source, the South Saskatchewan River flows for another 1392 kilometres. The SSR has an average discharge of 280 m3/s at the Saskatchewan River Forks. It covers a watershed of approximately 146,100 km<sup>2</sup> , of which 1800 km<sup>2</sup> are in Montana, United States, and 144,300 km<sup>2</sup> are in Alberta and Saskatchewan [29]. Agriculture accounts for two-thirds of the land cover [30].

### *2.2. Modelling Approach*

The change of heavy metals in the South Saskatchewan River system was studied using the Water Quality Analysis Simulation Program 7.52 (WASP) coupled with HEC-RAS. The WASP model could be used to simulate many different aquatic systems. The application could simulate hazardous water contamination using the concepts of mass, momentum, and the conservation of energy. The hydrodynamics of the model domain was simulated using the hydraulic model HEC-RAS 6.3. Since HEC-RAS was developed by the U.S. Army Corp of Engineers (USACE), all users and organizations have access to a free version of the model. HEC-RAS is available to everyone and does not require a license, making it a key benefit of utilizing it as a modelling tool. HEC-RAS can model flow, sediment transport, water quality in one-dimensional (1D) steady and unsteady flow, and two-dimensional (2D) unsteady flow of rivers. The model uses geometric data representation and geometric and hydraulic computer algorithms for a network of river channels. The HEC-RAS program can simulate an input flood using either a (1D) unsteady flow model or a (2D) unsteady flow model following these parameters.

Flow data were acquired from available flow data from the Water Survey of Canada (WSC) station at Lake Diefenbaker, which has an operational gauge (records available from 1966 to the present). That station is about 13 km upstream of the modelling upstream boundary (gauge 07DA001—South Saskatchewan River Below Lake Diefenbaker). The HEC-RAS model was calibrated by modifying Manning's N to obtain the best match of the water level at 05HG001—South Saskatchewan River at Saskatoon and 05KD007—Saskatchewan River below the forks.

A 1D modelling approach was chosen as the optimum method for simulating the South Saskatchewan River sites. A 1D modelling approach was found to be reasonable for the region with optimum processing times. One-dimensional modelling can give results that are on par with, or better than, 2D models for rivers and floodplains where the predominant flow directions and forces follow the main river flow route, with less effort and fewer computing resources [31]. It was expected that the system was mixed both laterally and vertically. Previously, 1D WASP and HEC-RAS coupling [32] and quasi-2D [33,34] have been carried out.

HEC-RAS carried out the hydrodynamic part of the modelling, and the output of the HEC-RAS model was used as an input (for water depth at each segment) for the WASP model.

#### *2.3. Setup for Water Quality Modelling*

The water quality part of the study was modelled using the Water Quality Analysis Simulation Program (WASP 7.32). Due to some stability problems, the newer version of WASP was not considered for the current study. WASP was first created in the 1980s and has undergone numerous improvements [35]. The general dynamic model WASP uses a segmentation network to solve the conservation of momentum, energy, and mass equations and simulate the transport of contaminants and sediment. The WASP model is frequently utilized to address sediment transport [34], heavy metal transfer [33], and water quality issues [36,37]. The WASP stream transport module, TOXI, is coupled with flow routing for free-flow streams, ponded segments, and backwater reaches and is capable of calculating the flow of water, sediment, and dissolved constituents across branched and ponded segments. Additionally, model boundary constraints and input parameters are defined.

Water flows through a network of branching streams, which may contain both freeflowing and ponded parts, and is calculated using the standard WASP8 stream transport model. Flow routing can be estimated for free-flowing stream reaches, ponded reaches, and backwater or tidally-influenced reaches for one-dimensional branching streams or rivers. Advective transport can be driven through free-flowing portions with the simple yet practical kinematic wave flow routing method. The kinematic wave equation determines the propagation of flow waves and the fluctuations in flows, volumes, depths, and velocities that arise from the varied upstream inflow. This well-known equation is the solution to the one-dimensional continuity equation and a simplified form of the momentum equation that considers gravity and friction.
