The fate of dissolved sulfamethoxazole during spring-thaw snowmelt in a field with a history of manure application

Abstract The fate of sulfamethoxazole (SMX) in Prairie agroecosystems during snowmelt is not well understood. This study aims to provide the first estimates of concentrations and loads of SMX in snowmelt in a field with a history of manure application. The mean concentration of SMX throughout the snowmelt period was 0.0345 ± 0.066 µg/L. The SMX cumulative load was 3.81 ± 3.4 µg/L with a range of 1.03–12.8 µg/L. Both the concentration and load were not influenced by the method of manure application (i.e., surface applied versus sub-surface applied).


Introduction
Sulfamethoxazole (SMX) is a broad-spectrum sulfonamide antibiotic used to treat bacterial infections in livestock operations (Kim et al. 2011).SMX is not completely metabolized by livestock, and up to 90% of SMX is excreted in manure as its original form (Kim et al. 2011).Application of liquid swine manure (LSM) has been found to lead to higher concentrations of both SMX and antibiotic-resistant genes (ARGs) in the soil compared with manure from other animals (Yue et al. 2021).This is especially concerning when the dissipation of sulfonamides has been shown to be slower in prairie wetlands (Cessna et al. 2020).Given the high rates of LSM use on agricultural fields in Manitoba, there is a pressing concern to understand SMX fate in the prairie environment and to develop management practices specific to this cold agricultural region.
The hydrological regime on Canadian Prairies is dominated by the spring snowmelt, which contributes up to 80% of the annual water runoff (Liu et al. 2018).Snowmelt can result in an extended period of flooding on agricultural fields and can potentially mobilize contaminants from soils to flood waters.While considerable attention has been given to the fate of nutrients during snowmelt (e.g., Costa et al. 2018), antibiotics have received less research.Furthermore, manure management efforts have primarily focused on preventing nutrient pollution in surface water (Liu et al. 2018).Research conducted during the growing season suggested incorporating surface-applied manure shortly after application can minimize the introduction of antibiotics into aquatic environments during rainstorms (Amarakoon et al. 2014).Comparable work has not been done during snowmelt when prolonged flooding and runoff present an increased risk of antibiotic transport.
This study aimed to quantify the concentrations and loads of SMX in a field with a history of manure application during snowmelt under different LSM management practices (no LSM applied, LSM applied on the surface, and LSM applied to the subsurface of soil).We hypothesized (i) SMX concentrations and loads will depend on the LSM application methods and (ii) the concentrations and loads of SMX will increase as snowmelt progresses when there is increased contact between snowmelt water and soil.

Experimental design and field sampling
In the Fall of 2021, 12 plots were established in an agricultural field that routinely receives manure applications based on the nutrient requirements and the soil test phosphorus levels in south-eastern Manitoba, Canada (49 • 32 5 N, 96 • 51 5 W).The soil at this site belongs to the Osborne series and is classified as Rego Humic Gleysol.The soil texture is clay, with a pH of 7.9 and organic matter content of 7.1%.We used a randomized complete block design with four replicate plots, where three treatments were applied to each block: LSM spread on the surface of the plot using small pails to evenly distribute the manure to ensure the accuracy of the manure rate (i.e., surface applied), LSM applied 15 cm into the soil, using a hoe to dig narrow trenches 15 cm deep which were then re-covered with soil (i.e., sub-surface applied), or no LSM applied (i.e., control).The LSM (∼2% solid content) was applied with an even distribution at an application rate of 170 000 L/ha, the manure application rate practiced by the producer in that year.The plots were 3 m × 1 m in size, with a 0.6 m alley between plots.Runoff boxes made of high-density polyethylene puckboard supported by wooden frames (1.2 m × 0.9 m × 0.6 m) were placed 15 cm into the soil surface to collect snow over the winter for each plot.
From 19 March to 5 April 2022, snowmelt samples were collected on days when the temperature was above 0 • C (between 12 pm and 2 pm), and liquid water was present on the surface of the soil.The average maximum temperature on sampling days was 4.2 • C, and the total precipitation over this period was 10.3 mm, including rain on days 3 and 7.Over the snowmelt period, there were 10 days where there was meltwater to collect.Floodwater in each runoff box was pumped out with a handheld water pump, and the total snowmelt volume per runoff box (i.e., replicate) was recorded, and a 250 mL subsample was taken from each plot.Samples were composited over two sampling days to ensure a sufficient mass of SMX for analysis from each subsample.

Sample preparation
The samples were pre-concentrated using solid-phase extraction (SPE).SPE Restek Hydrophilic Lipophilic Balance (HLB) cartridges (6 mg of sorbent, 3 mL; Oasis Waters) were conditioned, and the samples were passed through the HLB cartridges with a vacuum at a flow rate of 1 mL/min.Once all samples were pulled through the HLB cartridges, the cartridges were rinsed with 3 mL ultrapure water (Milli-Q; 18 M cm) to remove excess salts and allowed to dry for 1 min under vacuum.The samples were stored at −20 • C until they were transferred, 24 h before elution, to a refrigerator at 4 • C.
The samples were eluted with 3 mL of MeOH, allowing gravity to pull the samples through the HLB cartridges, followed by drying the samples to completeness using a 45 • C water bath and a nitrogen evaporator (OA-SYS Heating system and N-Evap 111, Organomation Associates, Inc.) with a gentle stream of Ultrapure N 2 (Praxiar Canada Inc.).The samples were reconstituted with 50/50 (v/v) MeOH: Milli-Q to 0.5 mL.Finally, the samples were filtered using a 0.22 μm syringe filter (Resteck) into 2 mL amber glass vials (Chromatographic Specialties Inc.).The samples were stored at -20 • C until analysis.

Liquid chromatography tandem mass spectrometry
Chromatography was performed with an Agilent 1260 Infinity II UHPLC (Agilent Technologies), with separation using an Agilent Eclipse Plus C 18 column (2.1 mm × 50 mm, 1.8 μm dp) coupled to an Agilent Eclipse Plus C 18 guard column (2.1 mm × 5 mm) at 42 • C at 0.3 mL/min.The injection volumes were 2 μL during optimization and 10 μL during analysis.Mobile phase A was Milli-Q water, and mobile phase B was MeOH.Gradient elution was performed as follows: 0-2.00 min linear ramp from 5% to 95% B, 2.01-4.00min hold at 95% B, 4.01-5.00linear ramp from 95% to 5% B, followed by re-equilibration from 5.01 to 8.00 min at 5% B.
Qualitative assessment and quantification were performed through multiple reaction monitoring on an Agilent 6470 triple quadrupole mass spectrometer in positive electrospray ionization mode (ESI+), a capillary voltage of 4000 V, and a source temperature of 300 • C. Nitrogen was used for desolvation and drying gas at 11 L/min, and for nebulization at 15 psi.Ultrapure nitrogen was used as collision gas at a flow of rate 16.8 L/min.The MS1 and MS2 heaters were set at 100 • C. The samples were spiked with deuterated SMX (Sigma-Aldrich) as the internal standard.The method limit of detection was 0.0094 μg/L, and the method limit of quantification was 0.015 μg/L.The % recovery of the method for SMX was 73.0 ± 38%.The linearity was R 2 = 0.9999.

Statistical analysis
A repeated measures ANOVA analysis was used to test the effects of manure application method and sampling day on SMX concentration.The treatment and sampling day were fixed effects, with block as a random effect, and the sampling day was the repeated factor.
The cumulative load of SMX was calculated by multiplying the SMX concentration by the mean daily snowmelt volume to account for high variance in snowmelt volume among replicates.A one-way ANOVA was used to test the effects of manure application method on the cumulative load of SMX.
The data were log-transformed when residuals were not normally distributed.All statistical analyses were performed in R (version 4.2.2,R Development Core Team 2022).

Results
The mean SMX concentration over the entire sampling period of all treatments was 0.0345 ± 0.066 μg/L and did not vary significantly among treatments (i.e., LSM application method; F-calculated = 0.223, p = 0.805; Table 1) or sampling days (F = 1.023, p = 0.410; Table 1).The maximum concentration introduced into the environment was 0.36 μg/L.Snowmelt volumes collected were highly variable among plots (i.e., replicates), as is expected on the prairies where wind redistribution of snow is widespread (Costa et al. 2018).The volumes also varied throughout the sampling period; the maximum volumes generally occurred on day 3 (37.1 ± 6.5 L) and went as low as 8.8 L by the final sampling day.The mean cumulative load of SMX was 4.12 ± 3.6 ng/m 2 for the 17-day snowmelt period, with a range of 1.03-12.8ng/m 2 , and there was no significant difference among the manure application method (F = 0.241, p = 0.787; Table 1).

Discussion
Sulfonamides pose a risk to the ecosystem health of freshwaters and are also linked to the rise in antibiotic resistance in environmental bacteria (Yue et al. 2021), but research on SMX transport dynamics from agricultural soils is limited.
Table 1.The least-square mean concentration (μg/L) and cumulative load (ng/m 2 ) with the standard error (in brackets) of sulfamethoxazole in snowmelt for the treatments (control, sub-surface applied, and surface applied liquid swine manure) and the sampling day (1, 3, 5, 13, and 17).In the Canadian Prairies, where most runoff comes during snowmelt, this study provides baseline data on the behaviour of this contaminant during snowmelt.There were no significant differences in SMX concentration among the two manure application methods or the control treatments where LSM was not applied in Fall 2021 (Table 1).The presence of SMX in the snowmelt from control plots could be due to the long history of manure application at this site with elevated residual SMX levels.This also meant that there were no significant differences in the cumulative loads among treatments, although there was a peak in cumulative load during days 3-5 of the sampling period, coincident with the largest volume of snow (Figure 1).While it was likely that the soil was still partially frozen on days 3-5, there was sufficient contact between the snowmelt water and the shallow soil to mobilize some SMX (Costa et al. 2018).This suggests that a high risk of transport of SMX into surface waters occurs when the snowmelt volume is highest.
The concentration of SMX required for chronic ecotoxicity in aquatic species ranges from 5.3-253 000 μg/L (Straub 2016).The concentrations of SMX in the snowmelt were below those levels, but the mean cumulative load of 4.12 ng/m 2 still poses a risk to aquatic systems, considering the widespread land application of manure (Liu et al. 2018).Another risk of SMX pollution in aquatic systems is the possibility of increasing antibiotic-resistant bacteria.The prolonged application of swine manure has been shown to result in higher levels of ARGs relative to other types of livestock manure (Wu et al. 2022).Thus, even if SMX concentrations in snowmelt are below ecotoxicological levels, there are environmental risks associated with snowmelt-induced loading in prairie environments.
Snowmelt volume primarily drives the cumulative load of SMX over the snowmelt period since there were no significant differences in SMX concentrations among treatments or days.The cumulative load continued to increase until the end of the snowmelt, albeit at a slower rate in the last 4 days (Figure 1).Management practices that hold snowmelt runoff on the land (Liu et al. 2018) may help to mitigate the risk of SMX pollution in streams and lakes.Future research should investigate the efficacy of such management options, which may also mitigate nutrient pollution (Liu et al. 2018).

Conclusion
Mitigating the transport of antibiotics into soils and surface waters is important to reduce adverse impacts on the ecosystem and human health.This study is among the first to quantify SMX in snowmelt water from manure-amended agricultural land under field conditions, and as such, provides data on possible concentrations (upto 0.36 μg/L) and load (1.03-12.8ng/m 2 ).This research will guide future research in this understudied area of antimicrobial transport in the Canadian Prairies to assist multi-faceted approaches to manure management for sustainable livestock cropping systems.

Fig. 1 .
Fig. 1.Boxplot of the cumulative load (ng/m 2 ) of sulfamethoxazole in snowmelt during the snowmelt period in three different treatments: (A) control, (B) sub-surface applied liquid swine manure, and (C) surface applied liquid swine manure in the Fall of 2021.The samples were collected over 17 days in the spring of 2022.