Loss of potentially toxic elements to snowmelt runoff from soils amended with alum, gypsum, and Epsom salt

Abstract Soil amendment effects on the mobility of potentially toxic elements (PTEs) have been hardly investigated under snowmelt flooding conditions. This research quantifies and compares the loadings of arsenic (As), copper (Cu), nickel (Ni), selenium (Se), vanadium (V), and zinc (Zn) to snowmelt from unamended, alum-, gypsum-, and Epsom salt-amended soils from a manured agricultural field and a non-manured agricultural field. In the fall of 2020, amendments were surface applied at a rate of 2.5 Mg ha−1 to field plots with four replicates. Runoff boxes were installed at the plots’ edge to collect winter snow. In the spring of 2021, the snowmelt in each box was pumped out, and volume was recorded until all snow in the boxes had melted. Concentrations of PTE and other cations and pH were measured in a subsample of the snowmelt. The snowmelt from the manured field had higher Ni, Se, and V loads than that from the non-manured field. There were no significant differences in snowmelt PTE loads between the amended soils and the unamended controls at each field. Although not statistically significant, the Epsom salt-amended treatment resulted in a 75% reduction in Se loading and a 44% reduction in V loading, while the gypsum-amended treatment showed a 38% reduction in Ni loading compared to the unamended treatment in the manured soil. Overall, our findings from a single season using both manured and non-manured fields suggest that alum, gypsum, and Epsom salt additions did not significantly alter the mobility of the studied PTEs during the spring snowmelt period.


Introduction
The loss of potentially toxic elements (PTEs) from agricultural fields during runoff poses an environmental threat to nearby water bodies, impacting the quality of drinking and irrigation water as well as aquatic life (Duodu et al. 2016;Simmatis et al. 2022).Additionally, the accumulation of PTEs in neighboring fields from agricultural runoff gradually increases soil toxicity, reducing its suitability for sustainable agricultural production (Wang et al. 2023).The PTEs are introduced to agricultural fields from a combination of anthropogenic and natural processes (Imseng et al. 2019;Liu et al. 2023).The mobility of PTEs in the soil is influenced by geochemical properties such as pH, redox potential, organic matter content, cation exchange capacity (CEC), and mineralogical composition (De Matos et al. 2001;Indraratne et al. 2023).To improve soil quality and reduce the mobility of PTEs and phosphorus, various amendments can be applied (Lizarralde et al. 2021;Indraratne et al. 2023;Kong and Lu 2023).However, these amendments can modify soil geochemical properties, increasing the bioavailability and mobility of some PTEs (Lwin et al. 2018).
Alum (aluminum sulfate), gypsum (calcium sulfate), and Epsom salt (magnesium sulfate) are commonly used soil amendments to improve soil structure, provide nutrients, alter pH, and reduce phosphorus mobility in soils (Ikoyi et al. 2020;Kumaragamage et al. 2022;Vitharana et al. 2021;Xu et al. 2023).Alum has demonstrated some capacity to adsorb selenium (Se), mercury (Hg), arsenic (As), and copper (Cu), but its effectiveness can be influenced by soil characteristics and environmental conditions (Ippolito et al. 2009).Gypsum acts as an immobilizing agent for Cu, chromium (Cr), cadmium (Cd), lead (Pb), and As (Lwin et al. 2018).Even at relatively low concentrations (1%, w/w), gypsum addition to contaminated soil/red mud mixtures enhanced aluminium (Al), As, and vanadium (V) adsorption and inhibited their mobility (Lehoux et al. 2013).However, the addition of these amendments can also change soil geochemical parameters, including Eh, pH, and dissolved organic carbon, thereby influencing the mobility of certain PTEs in soils (Shaheen et al. 2014;Indraratne and Kumaragamage 2018).In a simulated summer flooding experiment, soil amendment zeolite enhanced the mobility of V concentrations in floodwater (Indraratne et al. 2023).Dubrovina et al. (2021) found that adding gypsum to copper-smelter-polluted soils increased the concentrations of soluble Cd, manganese (Mn), and Pb, thereby enhancing metal uptake by plants.
In the Red River Basin flat landscape in Manitoba, Canada, nutrient loss from agricultural fields and the contamination of nearby water bodies primarily occur through snowmelt runoff (Rattan et al. 2017).Consequently, snowmelt runoff can transport various pollutants, including PTEs such as As, Cu, Ni, Se, V, and zinc (Zn), from agricultural fields to water bodies.However, the impact of amendments on the mobility of PTEs during snowmelt runoff has received limited attention in previous studies.To address this knowledge gap, our research aims to quantify and compare the levels of As, Cu, Ni, Se, V, and Zn in snowmelt from soils that were either unamended or amended with alum, gypsum, or Epsom salt.We collected these soils from both manured and nonmanured agricultural fields.We hypothesized that the addition of alum, gypsum, and Epsom salt amendments to manured and non-manured soils may affect the mobility of PTEs differently during the spring snowmelt.The specific objectives are to compare the loadings of PTEs (As, Cd, Cu, Pb, Ni, Se, V, and Zn) in spring snowmelt and potentially available fractions of PTEs in soils after the snowmelt period in alum-, gypsum-, and Epsom salt-amended plots with unamended plots in a manured field and a non-manured field, separately.

Properties of soils and amendments
Two agricultural fields were chosen from the Red River Basin in Manitoba, Canada, based on their identification as phosphorus (P) hotspots using data from the Lake Winnipeg Basin Community-Based Monitoring program (https://datastream.org/dataset/f 10bb610-63cc-46c1-81 b1-74a6b0310655).One of the fields had a known manure history (manured soil) and belonged to the Osborne Soil Series (Rego Humic Gleysol), and the other was a non-manured field from Red River soil series (Gleyed Rego Black Chernozem).The manured soil received liquid swine manure at a rate of about 140 000 L ha −1 (≈2% solid content) 3 days before the study setup.Four random soil samples (0-15 cm) were collected from each field and composited for the analysis of basic soil properties.Soil samples were analyzed for organic matter by loss on ignition (Dean 1974), pH (1:2 soil-water), particle size analysis (Gee and Bauder 1986), and cation exchange capacity (sum of exchangeable calcium (Ca), magnesium (Mg), sodium (Na), and potassium (K) using ammonium acetate method).
The total concentrations of a large number of elements including As, Cu, Ni, Se, V, and Zn in the amendments (alum, gypsum, and Epsom salt) as well as in the soil samples (manured and non-manured) were determined through a onestep digestion process.For this analysis, 0.25 g of soil or material was digested using a mixture of perchloric, nitric, and hydrofluoric acids.The digested samples were then analyzed using inductively coupled plasma mass spectrometry (ICP-MS; Agilent 7900) at ALS Canada Ltd. (method code ME-MS61).
To ensure accurate results, the analytical data were adjusted for interelement spectral interferences following the quality control procedures outlined in the QC certificate provided by ALS Canada Ltd. Mehlich-extractable PTEs from two soils were also determined as explained below.
In addition, total PTEs in a manure sample were determined by drying at 60 • C and digesting triplicate samples with sulfuric acid (H 2 SO 4 ) and hydrogen peroxide (H 2 O 2 ) at 350 • C for 3h.The acid-digested manure samples were diluted using deionized water with a conductivity of 18 M cm (Milli-Q water) and analyzed for PTE concentrations using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Thermo iCAP 6500 Duo, Cambridge, UK).

Field experiment
The experiment was conducted from the fall to spring of 2020-2021; amendments were added in the fall of 2020, and snowmelt water was collected in the spring of 2021.The experiment had four treatments (three soil amendments and an unamended control) with four replicates set up as a randomized complete block design.The manured and non-manured sites were seeded with wheat and canola, respectively, in the summer of 2020 prior to the establishment of this study.Soil amendments used were alum (Al 2 (SO 4 ) 3 •18H 2 O), gypsum (CaSO 4 •2H 2 O), and, Epsom salt (MgSO 4 •7H 2 O).In the fall of 2020, plots were set up along the field's edge.Each plot measured 3 m × 1 m, with a 0.5 m gap between plots.Runoff boxes (1.2 m long, 0.9 m wide, and 0.6 m high) were installed about 5 cm below the soil surface in each plot (Fig. 1).Six days after the installation of runoff boxes, soil amendments were applied to the runoff box of each plot at a rate of 2.5 Mg ha −1 .The amendments were mixed into the top 5 cm of the soil in all the plots and were left in place over the winter to collect snow.The amendment rate of 2.5 Mg ha −1 was decided based on previous studies conducted with P-sorbing amendments (Vitharana et al. 2021;Kumaragamage et al. 2022;Lasisi et al. 2023a) to the plots.During the spring snowmelt period in 2021, we visited the fields daily from 12:00 p.m. to 2:00 p.m. Melted snow was pumped out from each runoff box on each day it was present and the total volume of daily snowmelt was recorded.Meltwater samples were collected on four days (4, 5, 6, and, 7, March 2021), in non-manured soil and for six days (4, 5, 6, 7, 9, and, 13, March 2021) in manured soil.Details of the weather conditions during the study period (fall 2020 to spring 2021), snowmelt collection procedure, precautions taken, and observations made during the collections were given in detail in Lasisi et al. (2023b).Snowfall was lower in the 2020/2021 winter than the average yearly snowfall, and wind created a very high variability in the amount of snow collected in boxes (Lasisi et al. 2023b).

Elemental analysis in snowmelt samples
Snowmelt samples were syringe filtered (0.45 μm) within 6 h of collection, and samples were acidified with 50 μL of concentrated nitric acid (trace metal grade) and stored at 4 • C until analyzed for concentrations of Ca, Mg, iron (Fe), Mn, Al, S, P, and PTEs (As, Cu, Ni, Se, V, and Zn) using ICP-AES spectroscopy (Thermo iCAP 6500 Duo, Cambridge, UK).Selection of the PTEs was made based on the findings of our previous study (Indraratne and Kumaragamage 2018) and the total elemental composition of the two soils.The detection limits of ICP-AES for PTEs were 0.0030, 0.0005, 0.0011, 0.0015, 0.0008, and 0.0004 mg L −1 for As, Cu, Ni, Se, V, and Zn, respectively.Individual daily metal loads released to the snowmelt were calculated as the product of As, Cu, Ni, Se, V, or Zn concentrations (mg L −1 ), respectively, in snowmelt on a given sampling date multiplied by the snowmelt volume (L) measured on the same date.Cumulative PTE loads were calculated by summing daily loads during the sampling period, while cumulative snowmelt volume was calculated as the sum of daily snowmelt volume collected during the sampling period.The individual PTE load was converted to g ha −1 , using the area of the runoff box.

Mehlich-extractable PTEs in soils
Soil samples were collected (in duplicates) from all treatments at the end of the snowmelt period from each runoff box at 0-10 cm depth with a back-saver probe (Clements Associates Inc., IA, USA) and were composited to form one sample per runoff box.These samples were dried, ground, sieved (<2 mm), and analyzed for As, Cu, Ni, Se, V, and Zn extracted from Mehlich-3 solution (0.2 N CH 3 COOH-0.25N NH 4 NO 3 -0.015N NH 4 F-0.013N HN0 3 -0.001mol/L EDTA; Mehlich 2008) using ICP-AES spectroscopy (Thermo iCAP 6500 Duo, Cambridge, UK).

Statistical analysis
Analysis of variance (ANOVA) for PTE loadings to snowmelt was performed using the generalized linear mixed model (GLIMMIX) in SAS (SAS Institute Inc. 2013).The fixed effects were the treatment (i.e., three amendments and unamended control), and the site, with replicates nested within the site (manured and non-manured) as the random effect.For all the treatments, the elemental concentrations in meltwater were averaged across the sampling dates.These data were subjected to ANOVA with the treatment as the fixed effect and the replicates nested within the field site as the random effect.The Tukey multiple comparison procedure was used for all pairwise comparisons of treatment mean.When PTE concentrations were below detection limits, a value corresponding to half of the detection limit of the element was as-signed before analysis (Wong et al. 2002).Residuals of all data were checked for conformity to the assumption of normality.Non-normally distributed data were lognormally transformed during the ANOVA to meet the assumption of normality.

Results and discussion
Soil, manure, amendment, and snow characteristics Both non-manured and manured soils had a clay texture, with a neutral pH (7.5 and 7.7, respectively).The organic matter content was 6.5% and 7.5%, and the CEC was 36 and 70 cmol(+) kg −1 for the non-manured and manured soils, respectively (Table 1).
Concentrations of As, Ni, and V in the manure were very low (<11 mg kg −1 ), suggesting that the contribution of manure to total As, Ni and, V in manured soil is negligible.Total Se and Zn in liquid manure were 3685 and 1363 mg kg −1 , respectively, indicating the possibility of increasing total concentrations in manured soil with successive manure application; however, both manured and non-manured soils had similar total Se concentrations of ∼1 mg kg −1 (Table 1).
The contribution of PTEs from amendments was <3 mg kg −1 except for Epsom salt, which contributed 8 mg kg −1 of Cu and Zn, and gypsum contributed 20 mg kg −1 of Zn.The element contributions to soils due to the addition of alum, gypsum, and Epsom salt were not substantial based on the total elemental composition of amendments and the rate of application (Table 2).
Initial concentrations of Mehlich-extractable As, V, and Zn were 51%, 23%, and, 44% higher, respectively, in the manured than the non-manured soil.On the other hand, the Mehlichextractable Ni concentration in the non-manured soil was 55% higher than that in the manured soil, while both the manured and non-manured soils had similar concentrations of Mehlich-extractable Cu and Se (Table 1).
The total concentrations of As, Ni, and V were 82%, 38% and, 43% greater, respectively, in the non-manured soil than the manured soil.The differences between two soils were substantially less for Cu (37 and 33 mg kg −1 ) and Se (∼1 mg kg −1 in both soils).In contrast, total Zn was 19% higher in manured soil compared to non-manured soil (Table 1).Total PTE concentrations were below the soil quality guidelines (As, Cu, Ni, Se, and Zn concentrations of 12, 63, 45, 1, and 250 mg kg −1 , respectively) for agricultural soils in both manured and nonmanured soils (CCME 2006).Total V in non-manured soil was higher than the CCME threshold value for V in soil, which is 130 mg kg −1 (CCME 2006).
The proportions of total and Mehlich-extractable concentrations of the PTEs were similar between manured and nonmanured soils, except for As and V.This indicates that although the total concentrations of As and V were higher in the non-manured soil than the manured soil, the labile fraction of these elements is greater in the manured soils.Thus, despite having lower total concentrations, the potential mobility of As and V was higher in the manured soil.The higher P in manured soil could have influenced the mobility of As and V due to competition for adsorption sites on soil colloids (Cui and Weng 2013).Some of the elemental differences between manured and non-manured soils are probably due to geogenic differences in soils, since mineralogical and chemical characteristics of bedrock, texture, and local soil ahydrological conditions have been reported to affect the concentration and distribution of trace elements in agricultural soils of southern Manitoba (Haluschak et al. 1998).A previous study conducted using soils from the same region reported a similar variability of total As and V concentrations, as report in this study, ranging from 8.3 to 10.9 mg kg −1 , and 89 to 167 mg kg −1 , respectively, among four soils (Indraratne et al. 2023).It is worth noting that the soluble, exchangeable, and chelated species of PTEs in the soils are the labile fractions that can easily become mobile, even though total PTE concentrations are useful for determining soil quality and the potential risk of pollution (Buccolieri et al. 2010).
At each site, the As, Cu, Ni, Se, and Zn concentrations of the snow were <1 μg L −1 , indicating that the contribution from atmospheric deposition is not a substantial source of these PTEs (data not shown).The V concentration of the fallen snow was 5 μg L −1 , indicating the possibility of atmospheric deposition.Since V has the highest atmospheric anthropogenic enrichment factor of any trace metal (He et al. 2020;Kika et al. 2023), it is expected to have higher V concentrations in snow than other PTEs.

Potentially toxic elements in snowmelt
The snowmelt volumes were presented and discussed in detail in Lasisi et al. (2023b).In brief, the cumulative snowmelt volume ranged from 11.9 L (Epsom salt-amended) to 14.8 L (alum-amended) in non-manured soil and the treatment effect on the snowmelt volume was statistically non-significant.In manured soil, the cumulative snowmelt volume was significantly different only between the unamended control and the Epsom salt-amended treatments; unamended had the highest (31.1 L), followed by gypsum-amended (18.7 L), alumamended (16.7 L), and Epsom salt-amended with the least (14.4 L).Therefore, Epsom salt treatment had 2.2 times lower volume than the control (Lasisi et al. 2023b).Based on the nearest Weather Canada stations to the study sites (https: //climate.weather.gc.ca/), the study year (2020-2021) experienced severe drought conditions with a snowfall (October-March) of 19.5 cm (17% of climate normal during this period) at the non-manured site and 39.5 cm (42% of climate normal) at the manured site.The site effect (manured and non-manured) and treatment effect (amendments) in snowmelt PTE concentrations averaged across sampling days are given in Table 3.The snowmelt had <1 μg L −1 As concentrations in the manured and nonmanured soil and were below the critical thresholds (<100 μg L −1 ) for irrigation water (CCME 2006).Average Cu concentrations in snowmelt varied from 4.3 to 10.3 μg L −1 .The Ni concentrations (1.9-3.2 μg L −1 in manured soils 1.5-2 μg L −1 in non-manured soils) in snowmelt were below the irrigation water quality guidelines (200 μg L −1 ; CCME 2006).Average Se concentrations in snowmelt of non-manured and manured soils varied from 1.3 to 4.7 and 4.3 to 7.7 μg L −1 , respectively, and these concentrations were higher than the CCME (2006) guidelines imposed for the protection of aquatic life (1 μg L −1 ).Snowmelt V concentrations were well below the irrigation water quality threshold (100 μg L − 1; CCME 2006) in both manured (18.1-20.0μg L −1 ) and non-manured (3.5-4.1 μg L −1 ) soils, but exceeded the threshold for sensitive aquatic species (2-80 μg L −1 ; Abernathy et al. 2021) and drinking water standard (15 μg L −1 ; Schlesinger et al. 2018), especially in manured soils.Average Zn concentrations varied between 18.8 and 32.8 μg L −1 among treatments of manured and non-manured soils.Significantly higher concentrations of As and Zn were observed in the snowmelt from non-manured site compared to manured site.Conversely, snowmelt in manured site showed significantly higher concentrations of Se, V, and Ni compared to non-manured site.There were no significant differences in snowmelt Cu concentrations between the two sites.There were no significant differences among treatments in both manured and nonmanured soils for As, Ni, Se, V, and Zn.Therefore, the addition of alum, gypsum, or Epsom salt did not increase PTEs (except Cu) in the snowmelt in both manured and non-manured soils.In contrast, the amendment effect for Cu was significant among treatments, indicating significantly higher Cu concentration in the snowmelt of the alum than that of Epsom salt.However, snowmelt Cu in unamended or gypsum-amended treatments was not significantly different than that of alum or Epsom salt treatments.This finding aligns with a previous study by Lombi et al. (2010), which demonstrated that alum amendment for poultry litter resulted in increased solubility of Cu.
The snowmelt volume was recognized as one of the main drivers influencing the phosphorus loads from snowmelt runoff at these fields (Lasisi et al. 2023b), and in other fields used for snowmelt studies in the Canadian prairie (Wilson et al. 2019).Hence, we calculated PTE loads to determine site (manured and non-manured) and amendment (alum, gypsum, Epsom salt amended and unamended) effects on the PTE loadings to surface water (Fig. 2).The manured and nonmanured field effect of As, Cu and Zn loadings in snowmelt were non-significant, while Se, V, and nickel (Ni) loadings in snowmelt were significantly higher in manured field than the non-manured field.The snowmelt loadings ranged from 0.03 to 0.07 g ha −1 for As, from 0.41 to 0.76 g ha −1 for Cu, and from 2.5 to 4.1 g ha −1 for Zn in both manured and non-manured fields.On average, the snowmelt loads of Se, V, and, Ni in the manured field were 69%, 86%, and 62% higher, respectively, than those in the non-manured field.The greater V loadings to snowmelt in the manured soil, despite the very low concentration of V in the manure itself (Table 1), may be due to several reasons.Manured soil had higher pH, organic matter content, and CEC compared to non-manured soil, indicating a higher buffer capacity against changes in pH.Snowmelt pH was lowest on the first sampling day in both soils (6.5-6.6 in the non-manured soils; 6.9-7.1 in the manured soils) and increased with sampling day, to an average of 7.4 in non-manured treatments and to an average of 7.8 in manured treatments (Lasisi et al. 2023b).Alkaline pH and organic C enhanced V mobility from soils (Shaheen et al. 2014;Shaheen and Rinklebe 2018).In a previous study, V mobility was increased with added zeolite amendment due to the increase of soil pH (Indraratne et al. 2023).Apart from pH changes, V transformation could also play a role in V mobility.Vanadium exists in several species, and among them V 5+ species are the most soluble form while V 4+ species are easily converted to an insoluble hydroxide, and complex with organic matter allowing it to remain stable over a neutral to alkaline range of pH (Wright et al. 2014;Huang et al. 2015).It was found that the V 5+ serves as an electron acceptor and the presence of other electron acceptors such as NO 3 − had the potential to interfere with V 5+ reduction (Liu et al. 2017b;Wang et al. 2020).In this study, we speculate that the presence of higher concentrations of NO 3 − in the manured soil interfered with the reduction of V 5+ to V 4+ resulting in higher V 5+ (mobile species) in manured soil than that of the non-manured soil.Therefore, irrespective of higher total V content in nonmanured soil (Table 1), the manured soil had significantly higher V load in the snowmelt.Total Ni concentration was higher in the non-manured soil compared to the manured soil.Yet, the Ni load in the manured snowmelt was higher than that of non-manured snowmelt.Ni mobility could be governed by enhanced bacterial richness index in bulk and rhizosphere soils because of the addition of manure.Meng et al. (2023) observed a correlation between availability of elements and compositional variations of bacteria and fungi in bulk and rhizosphere soils after addition of swine manure.
The amendment effect was not significant for snowmelt PTE loadings in both manured and non-manured sites.Al-though As loads among amended treatments in this study were not significantly different from the unamended control plots, a previous study showed a 38% of reduction of As bioavailability in contaminated paddy soils by titanium gypsum (Hussain et al. 2021).Among amendment treatments, the Epsom salt treatment had only 25% of snowmelt Se load (75% reduction) compared to the unamended control in the manured site, suggesting that while the amendments used in the study did not increase the mobility of Se, the Epsom salt treatment may decrease the mobility.A previous study reported that the application of elemental S or magnesium sulfate to soils reduced Se concentration in rapeseed (Liu et al. 2017a).Even though the treatment effect was not significant, similar to the results for Se, Epsom salt reduced 44% of V load in snowmelt compared to the respective unamended control in the manured site.Although the amended treatments did not show significant differences from the control treatment, gypsum also reduced Ni loading in the snowmelt by 38% compared to the unamended control in the manured soil.

Soil available PTE concentrations after snowmelt
Mehlich-3-extractable As, Cu, Ni, Se, V, and Zn from amended and unamended soils in manured and nonmanured soil sampled after the snowmelt period are shown in Fig. 3.The Mehlich-3 extraction method enables the extraction of a wide spectrum of elements, such as macro (e.g., Ca, Mg, K, and P), micro (e.g., Cu, Fe, Mn, Ni, and Zn), and toxic elements (e.g., As, Cd, Cr, and Pb) from weaklyto-strongly held pools in soils and considered a good indicator of the bioavailable fraction (Mehlich 2008;Han et al. 2020).Since the water-soluble fractions of PTEs were removed from soils to snowmelt, Mehlich-3-extractions indicated the potential mobile fraction to plant uptake and leaching losses.Mehlich-3-extractable As, Se, V, and Zn were higher in the manured soil than in the non-manured soil, while the differences among treatments were not-significant.Mehlich-3 ex-tractable Cu concentrations were not significantly different among treatments in both manured and non-manured soils.
In contrast, Mehlich-extractable Ni in the non-manured soils was significantly higher (2.85-3.09mg kg −1 ) than that in the manured soil (1.23-1.36mg kg −1 ), and there were no significant differences among treatments.Soil analysis showed that non-manured soil had higher total Ni concentrations than the manured soil.Although snowmelt Ni load in manured soil (water-soluble Ni) was greater than that in non-manured soil, the manured soil had a lower total and Mehlich-extractable Ni than the non-manured soil.The presence of high organic C and CEC in manured soil may play a role in retaining Ni more strongly.The combination of organic C and CEC in manured soil likely facilitates the binding and retention of Ni, reducing its availability for extraction by Mehlich-3 solution.In remediation studies, it has been consistently observed that functionalized biochars exhibit higher efficiency in removing Ni from contaminated soil compared to pristine biochars (El-Naggar et al. 2021).This enhanced removal can be attributed to the ability of functionalized biochars to retain the weakly-to-strongly held fraction of Ni in the soil through mechanisms such as ion exchange and electrostatic attraction.This phenomenon highlights the potential of manured soil to retain Ni and potentially contribute to its remediation and immobilization in contaminated environments.
Similar to snowmelt V loads, Mehlich-3-extractable V concentrations also showed significant differences between manured and non-manured soils.This difference is again opposite of the total V concentrations in non-manured soil (165 mg kg −1 ) and in manured soils (115 mg kg −1 ).It can be speculated that the changes in intrinsic soil properties (e.g., higher pH, CEC, and organic C in the manured soil than in the nonmanured soil) and V species transformation differences (restriction in V 5+ to V 4+ transformation in the presence of NO 3 − ) may have played a substantial role in releasing mobile forms, as Mehlich-3-extractable fraction indicated the potential for becoming mobile.The bioreduction of V 5+ may have been prevented due to the competition with other electron acceptors in the soil, with oxygen and nitrate being the most prevalent (Li et al. 2022).This competition could have contributed to the release of V in mobile forms, in the manured site.On the other hand, the concentrations of Mehlich-3extractable PTEs did not show significant differences among treatments, indicating that the addition of amendments did not increase the potential mobility of PTEs.This suggests that the amendments did not have an adverse effect on the mobility of PTEs in both manured and non-manured soils.

Conclusions
The application of alum, gypsum, and, Epsom salt as soil amendments to mitigate phosphorus losses did not result in increased mobility of potentially toxic elements (PTEs) such as As, Cu, Ni, Se, V, and Zn with snowmelt in either manured or non-manured prairie soils.Although some significant differences in PTE behavior were observed between manured and non-manured fields, we cannot draw substantial conclusions from these differences due to the absence of replication at the field level.
The behavior of PTEs varied in response to the three different amendments, as indicated by the percentage reduction of individual PTEs.Based on this 1-year snowmelt study with one manured site and one non-manured site, we can conclude that the addition of alum, gypsum, or Epsom salt did not contribute to the mobility of As, Cu, Ni, Se, V, and Zn with meltwater irrespective of the presence or absence of manure.
Nonetheless, these three amendments remain viable options for mitigating phosphorus mobility in prairie region soils.Since the choice of amendment may influence the mobility of various PTEs in diverse ways, it is advisable to conduct a comprehensive assessment of PTE mobility changes before introducing a new amendment.Furthermore, to gain a more comprehensive understanding of PTE mobility with different amendments, it is essential to conduct multi-site, multi-year field studies.

Fig. 1 .
Fig. 1.Collection of snow during the winter using the runoff boxes installed at the edge of the field.

Fig. 2 .
Fig. 2. Loadings of the potentially toxic elements (As, Se, V, Cu, Zn and Ni) in snowmelt of manured and non-manured fields (level of statistical significance is p ≤ 0.05 for main effect means; the different uppercase letters indicate significant differences among sites; NS = non-significant; error bars are standard error of the mean).

Fig. 3 .
Fig. 3. Soil Mehlich-3-extractable potentially toxic elements (As, Se, V, Cu, Zn, and Ni) in manured and non-manured soils after snowmelt period (level of statistical significance is p ≤ 0.05 for main effect means; the different uppercase letters indicate significant difference of site effect; NS = not significant; error bars are standard error of the mean).

Table 1 .
Initial soil properties of the soils.

Table 2 .
Total element concentrations (mg kg −1 ) in soils, manure, and amendments used for the study.

Table 3 .
Main effect means of treatment and site on concentrations (μg L −1 ) of potentially toxic elements (averaged across the sampling date) in snowmelt water in manured and nonmanured soils (different letters indicate significance at p < 0.05).