Assessment of viral pathogen removal across wastewater treatment stages in a swine processing facility

Abstract

A swine abattoir wastewater treatment system in Queensland, Australia, was monitored over 6 months to assess the effectiveness of a pond‐based treatment system for the removal of viruses prior to irrigation reuse. Four viruses were tracked through the facility, from points of entry to the final treated effluent, to identify major sources of viral contamination and quantify viral reductions across treatment stages. The truck wash contributed the highest viral loads overall, reflecting the accumulation of porcine faecal materials on transport vehicles and its release during washing. Substantial viral loads were also detected from the lairage area and slaughter floor, consistent with direct contact with live animals and carcase processing. In contrast, the boning room contributed the lowest viral loads among sampled locations. These findings identified priority locations for targeted hygiene interventions and drainage management. The three‐pond system achieved substantial viral reductions, with concentrations decreasing consistently through the anaerobic, aerobic and storage ponds. By the final pond, viral concentrations were generally at or below detection limits. Overall reductions ranged from approximately 90%–99.9%, depending on virus type. Soil samples collected from irrigated areas tested negative for all viruses, consistent with non‐detectable levels in the applied effluent, indicating no measurable viral presence in the soil following land application under the conditions studied. Viability‐based quantitative polymerase chain reaction (qPCR) and reverse transcription quantitative polymerase chain reaction (RT‐qPCR) assays demonstrated that most detectable viral genomes in treated wastewater originated from non‐intact, likely non‐infectious particles. Intact viral particles were rarely detected following final treatment, even when viral genomes remained measurable, suggesting that conventional qPCR/RT‐qPCR quantification alone may overestimate the presence of infectious viruses in treated effluent. Overall, the pond‐based treatment system provided effective viral removal prior to irrigation reuse. Maintaining adequate retention times is recommended to sustain treatment performance and ensuring the safety of irrigation reuse. Strategic management of truck wash effluent was identified as an opportunity for further viral load reduction.

Wastewater generated by abattoirs contains high concentrations of organic matter, suspended solids, nutrients and diverse microbial communities due to activities such as live animal handling, cleaning and wash down, slaughter and carcase processing. ¹ , ² These processes introduce faecal material, blood, other body fluids and animal tissue residues into the waste stream. ³ , ⁴ As a result, the wastewater carries microorganisms of animal and human health relevance, including pathogenic bacteria, protozoa and viruses of livestock origin, some with zoonotic potential or significant importance to livestock health and trade. ⁵ , ⁶ , ⁷ , ⁸

Management of wastewater from the livestock and meat‐processing sector involves several disposal pathways, depending on facility design and regulatory requirements. Common approaches include discharge to municipal wastewater treatment plants (WWTPs) as regulated trade waste, land application via irrigation or fertigation systems and release to natural or constructed wetlands or surface waters after some level of on‐site treatment. ⁹ , ¹⁰ , ¹¹ If treatment is inadequate, these effluents can introduce pathogens into soils and waterways via runoff or infiltration, posing risks to ecological receptors and human health. ¹² , ¹³ , ¹⁴

Despite the environmental relevance of abattoir wastewaters, research on their viral composition and fate has been limited. ¹⁵ Evidence suggests that viruses often persist longer in the environment and can withstand conventional wastewater treatment more effectively than bacteria. ¹⁶ Swine‐associated wastewaters can contain a diverse suite of viruses, including production‐relevant pathogens such as porcine adenovirus (PAdV), porcine circovirus type 2 (PCV2) and porcine rotavirus C (PRoVC), as well as enteric bacteriophages like MS2, which are commonly used as viral surrogates for assessing wastewater treatment performance. ¹⁷ , ¹⁸ Detection of viral genomes in treated effluents suggests that existing treatment processes may not always achieve adequate viral removal, underscoring the importance of virus‐specific evaluations of treatment efficacy.

Abattoir wastewater treatment systems typically employ physical and biological processes, including anaerobic and aerobic ponds, settling or clarification stages and storage basins, to reduce organic loads and suspended solids. ¹⁹ However, virus removal can be variable and is influenced by operational and environmental factors such as hydraulic retention time (HRT), interactions with suspended or settleable solids and microbial activity. ²⁰ , ²¹ Additional pond conditions, including sunlight exposure, temperature, turbidity and ammonia concentrations, further affect viral persistence. ²² , ²³ , ²⁴ , ²⁵

Observed viral removal efficiencies in municipal WWTPs, typically around 1–3 log₁₀ in the absence of tertiary treatment or disinfection, are well documented. ²⁶ , ²⁷ , ²⁸ , ²⁹ However, it remains unclear whether these patterns are applicable to abattoir‐based systems. Unlike municipal wastewater, abattoir waste streams have substantially higher solids content, elevated nutrient loads and distinct viral communities dominated by livestock‐specific pathogens. These compositional differences may influence viral persistence, partitioning and removal dynamics, potentially resulting in viral reduction outcomes that differ from those observed in municipal WWTPs. ³⁰ , ³¹

The objectives of this study were to: (1) characterize the occurrence and distribution of a viral surrogate (MS2 phage) and selected swine‐associated viral pathogens representative of different virus types (PAdV, PCV2 and PRoVC) across four key operational wastewater sources: lairage area (LA), slaughter floor (SF), boning room (BR) and truck wash (TW) to identify major contributors to viral load entering the combined wastewater systems (CWS); (2) quantify viral concentrations in CWS and throughout a pond‐based treatment train comprising anaerobic, aerobic, and storage ponds (Ponds 1–3); (3) determine log₁₀ reduction values (LRVs) across sequential wastewater treatment stages; (4) assess correlations between MS2 and pathogenic viruses to evaluate its suitability as a process indicator and (5) apply PMAxx™ viability assays to distinguish intact from non‐intact viral particles and compare viability‐adjusted reductions with conventional qPCR/RT‐qPCR estimates, with the aim of improving understanding of viral persistence and survival in a swine abattoir wastewater treatment system.

Materials and methods

Study site and wastewater treatment plant configuration

Wastewater samples were collected from LA, SF, BR and TW (operational areas), from a full‐scale swine abattoir WWTP in Queensland, Australia. The facility generates high‐strength wastewater containing elevated organic matter, nutrients and suspended solids. Treatment is provided by a continuous‐flow, multi‐stage pond system designed for progressive reduction of organic and nutrient loads, with final effluent applied to crops via irrigation. The CWS is screened to remove gross solids before treatment in a series of anaerobic, aerobic and storage ponds operating under continuous inflow and outflow conditions (Figure 1). The system is not batch‐operated; instead, each pond functions at approximately steady state, where HRT reflects the average time wastewater resides in each pond based on the flow rate and effective volume. The anaerobic pond (P1; 18 ML) has a typical HRT of ~7 days, while the subsequent aerobic effluent pond (P2; 17.8 ML) has a typical HRT of ~13 days. These values represent mean residence times under normal operating flow conditions rather than filling or emptying cycles. HRT is reported solely to describe system design and temporal mixing characteristics of the ponds and was not used in log reduction calculations. The total system HRT from raw influent to final effluent is approximately ~20 days. All operations are conducted in accordance with the facility's environmental authority.

Figure 1.

Schematic representation of the abattoir wastewater treatment system and sampling locations. Wastewater generated from operational areas (LA, SF, BR, TW) is combined into a single waste stream and treated sequentially in a three‐pond stabilisation system comprising an anaerobic Pond 1 (P1), aerobic Pond 2 (P2) and wet‐weather storage Pond 3 (P3). Red circles indicate sampling points at each operational source, at the combined influent and within each treatment pond. Effluent from the final pond is reused for irrigation under suitable conditions.

Sample strategy and collection

Wastewater samples (n = 72) were collected from four operational areas (LA, SF, BR, TW) and four treatment stages (CWS, P1, P2, P3). Soil samples (n = 9) from the effluent irrigation (EI) area were collected concurrently. Sampling was conducted over a six‐month period (July 2025–February 2026) on nine occasions (10/7/2025, 07/08/2025, 21/08/2025, 03/10/2025, 31/10/2025, 24/11/2025, 30/01/2026, 16/02/2026 and 27/02/2026) to capture temporal variability in wastewater characteristics and treatment performance. At each event, 100–500 mL grab samples were collected during routine operations by trained staff in sterile polyethylene terephthalate (PET) bottles following standardized grab sampling procedures (APHA, 2017). Samples from P1 and P2 were collected at the pond outflow points, while other samples were collected from accessible fixed sampling locations within the respective wastewater streams. All samples were immediately placed in insulated coolers with ice packs, transported to the laboratory within 24–48 h, stored at 4°C on arrival and processed within 24–48 h for molecular analyses.

Virus concentration from wastewater samples

Endogenous viruses were concentrated using an adsorption‐extraction (AE) workflow, ³² which has been applied for wastewater and environmental water samples with demonstrated recovery efficiencies. ³² , ³³ Filtration volumes were adjusted to account for differences in suspended solids and to minimize the risk of membrane clogging: 10–50 mL for LA, 1–50 mL for KF, 50 mL for BR, 1–50 mL for TW, 10–100 mL for CWS, 50–200 mL for P1 and P2 and 100–200 mL for P3. Samples were filtered through 0.45‐μm electronegative HA membranes (Merck Millipore, Tokyo, Japan) using a magnetic filter funnel (Pall Corporation) connected to a vacuum flask (Merck Millipore). ³² Membranes were aseptically removed, rolled and transferred into 5‐mL bead‐beating tubes (Qiagen, Hilden, Germany) for nucleic acid extraction. For soil samples, 0.5–0.6 g wet weight was added to 2‐mL bead‐beating tubes containing 0.1‐mm glass beads.

Nucleic acid extraction and quality assessment

An exogenous extraction control (murine hepatitis virus, MHV; ~6100 ± 400 genome copies [GC]) was added to each bead‐beating tube containing the concentrated sample, membrane and lysis buffer to monitor nucleic acid extraction performance. Lysis was performed by adding 990 μL of PM1 buffer (Qiagen) and 10 μL of β‐mercaptoethanol (Sigma‐Aldrich). Samples were homogenized using a Precellys 24 tissue homogenizer at 9,000 rpm for three 15‐s cycles with 10‐s intervals (Bertin Technologies, France). Tubes were centrifuged at 4000 g for 5 min, and lysates were transferred to clean 1.5–2 mL tubes. Total nucleic acids were extracted using the QIAcube automated platform with the RNeasy PowerWater Kit (Qiagen). Extracts were eluted in 150 μL nuclease‐free water. Concentration and purity of the samples were assessed using a DeNovix DS‐11 Series Spectrophotometer/Fluorometer (Wilmington, DE, USA). Purity was evaluated using A _260/280 and A _260/230 ratios, with A _260/230 ≈ 2.0 considered acceptable for RNA and A _260/280 ≈ 1.80 considered acceptable for DNA (Thermo Fisher Scientific, Interpretation of Nucleic Acid 260/280 Ratios, NanoDrop Technical Bulletin T123). Extracts were stored at −20°C until molecular analysis.

Assessment of nucleic acid extraction efficiency

Nucleic acid extraction efficiency was assessed using MHV as an exogenous extraction control. MHV was selected as it is not expected to occur naturally in swine abattoir wastewater, allowing it to serve as a reliable indicator of extraction performance. MHV was quantified using the RT‐PCR assay described in the qPCR and RT‐qPCR assays section.

Nucleic acid extraction efficiency was calculated as:

$$\text{Extraction efficiency}\left(\%\right)=\frac{\text{Measured}\mathrm{MHV}\left(\mathrm{GC}\right)}{\text{Seeded}\mathrm{MHV}\left(\mathrm{GC}\right)}\times 100$$

The seeded concentration was ~6100 ± 400 GC/sample. All reactions were performed in triplicate, and the mean value was used to calculate extraction efficiency. This approach provided a quantitative QA/QC measure to verify consistent nucleic acid extraction performance across samples.

Assessment of matrix‐associated RT‐qPCR inhibition

Matrix‐associated inhibition was evaluated using the exogenous MHV extraction control. The Cq value of MHV in each wastewater extract was compared with that of an MHV control prepared in nuclease‐free water and analysed under identical RT‐qPCR conditions. The difference in Cq (ΔCq) was calculated as follows:

where Cq,_sample is the Cq obtained from the wastewater extract, and Cq_control is the Cq value of the MHV control. An increase in ΔCq >2 was considered indicative of reduced RT‐qPCR performance attributable to matrix effects, including potential PCR inhibition and/or reduced extraction efficiency. ³⁴ Samples exceeding this threshold were diluted (1:5 or 1:10) and re‐analysed to minimize matrix‐associated inhibition.

qPCR and RT‐qPCR assays

Primer and probe sequences, concentrations and thermal cycling conditions are summarized in Table ST1. MHV was quantified using the RT‐qPCR assay of Besselsen et al. ³⁵  Published assays were used to detect and quantify MS2, ³⁶ PAdV, ³⁷ PCV2, ³⁸ and PRoVC ³⁹ . qPCR assays for PAdV and PCV2 were performed in 20 μL reaction volumes using 2 × QuantiNova Probe PCR Master Mix (Qiagen). Each reaction contained 10 μL of Master Mix, forward and reverse primers, probe (final concentrations shown in ST1) and 5 μL of template nucleic acid. RT‐qPCR assays for MHV, MS2 phage and PRoVC were conducted in 20 μL volumes using TaqMan™ Fast Virus 1‐Step Master Mix (Applied Biosystems, California, USA). Each reaction contained 5 μL of Master Mix, appropriate primers and probes and 5 μL of nucleic acid template. Synthetic gBlocks were used for standard curves and as positive controls. All reactions were performed in triplicate on a Bio‐Rad CFX96 thermal cycler. Automated baseline and threshold settings were used, and three no‐template controls were included per run.

Viability (intact‐particle) qPCR and RT‐qPCR assays using PMAxx™

Viability‐qPCR (v‐qPCR) and v‐RT‐qPCR assays were performed to differentiate between total and potentially viable (membrane‐intact) viral particles, allowing estimation of the viable intact fraction in wastewater samples. Variable sample volumes from the CWS (10 mL), P1 (25 mL), P2 (25 mL) and P3 (50 mL) were analysed. Samples from the final three sampling events (30/01/2026, 16/02/2026 and 27/02/26) were selected from the total of nine events for this analysis. Each sample was mixed thoroughly, and respective aliquots (10, 25 or 50 mL) were then transferred into two Amicon® Ultra Centrifugal Filters (Ultracel 30 kDa MWCO, Millipore, Darmstadt, Germany). Concentration was performed by centrifugation at 4500 g and 10°C for up to two or more consecutive 20–30 min intervals or until the retentate volume was reduced to <200 μL. Concentrates from both filters were combined and transferred to a 1.5‐mL microcentrifuge tube. The combined concentrate was then divided equally into two aliquots: one subjected to PMAxx™ treatment (Biotium, Fremont, CA, USA) and the other left untreated to represent total nucleic acids.

For PMAxx™‐treated samples, one‐quarter of the sample volume of 5× Enhancer solution was added and mixed thoroughly, followed by addition of 1/100 of the sample volume of 10 μM PMAxx™ solution (e.g., for a 140 μL sample, 35 μL enhancer and 1.8 μL PMAxx™). Samples were incubated in the dark at room temperature for 20 min to enable dye penetration into compromised viral particles. Both treated and untreated samples were subsequently placed in a PMA‐Lite™ 2.0 LED photolysis device (Biotium) for 30 min with oscillation at 300 rpm in the dark to induce photoactivation and covalent binding of PMAxx™ to exposed nucleic acids. Tubes were gently mixed at 10‐min intervals during this incubation. Following photoactivation, samples were transferred into 2‐mL bead‐beating tubes from the AllPrep PowerViral DNA/RNA Kit (Qiagen). To each tube (i.e., PMAxx™‐treated and untreated wastewater samples), 10 μL of β‐mercaptoethanol and 800 μL of Qiagen PM1 buffer were added to achieve a final volume of 1 mL. Samples were homogenized using a Precellys homogenizer at 10,000 rpm for three 15‐s cycles with 10‐s intervals (Bertin Technologies). Homogenates were centrifuged at 13,000 g for 1 min at room temperature, and supernatants were transferred to clean 2‐mL tubes.

Subsequently, 200 μL of IRS solution (Qiagen) was added to each tube, vortexed and incubated at 4°C for 5 min to reduce inhibitor carryover. After a 1‐min centrifugation at 13,000 g, the supernatants were transferred to fresh 2‐mL tubes. Aliquots of 375 μL were processed using the QIAcube Connect automated platform (Qiagen) following the manufacturer's protocol, without DNase digestion. Nucleic acids were eluted in two sequential steps (80 μL followed by 70 μL) into a single collection tube, yielding a final volume of 150 μL.

Data handling, transformation and statistical analysis

Viral concentrations were normalized to 100 mL to enable comparison across sample types. Log₁₀ transformation was applied to stabilize variance and facilitate estimation of reductions between treatment stages. Samples in which the target was not detected (non‐detects; NDs) were assigned a proxy value of 5 GC/reaction, and sample limit of detection (SLOD) values were calculated by incorporating initial volume, concentration factors and elution volume. This approach enabled inclusion of measurements below the assay detection limit (i.e., left‐censored observations, where the true concentration is unknown but falls below a defined threshold) in quantitative analyses while avoiding overestimation of treatment performance. ND values are reported as SLOD for descriptive purposes only and were treated as left‐censored in statistical analyses.

Log₁₀ reduction values (LRVs) were calculated as the difference between log₁₀‐transformed concentrations at sequential treatment stages, with overall treatment performance assessed from CWS to P3. Because the dataset included censored values and was not normally distributed, non‐parametric statistics were used. Spearman's rank correlation coefficients (r _s) were calculated to assess associations between MS2 and target viral pathogens across operational and effluent samples. Correlation analysis was undertaken to assess the suitability of MS2 bacteriophage as a process indicator by evaluating the extent to which it reflects the behaviour of selected swine‐associated viral pathogens across wastewater sources and treatment stages. Statistical analyses were performed using GraphPad Prism v11.0.0, with significance assessed at α = 0.05.

Results

Quality assurance and quality control (QA/QC)

Nucleic acid samples were successfully extracted from wastewaters across all operational areas and treatment stages and soil samples, yielding concentrations ranging from 2.9 to 3293 ng/μL for DNA and 1.3–2669 ng/μL for RNA (Table ST2). Most wastewater samples, particularly from LA, SF, TW, CWS and pond stages (P1–P3), exhibited high nucleic acid yields and met spectrophotometric purity criteria (A _260/280 = 1.90–2.10; A _260/230 = 2.00–2.40), confirming extraction consistency across sampling events. In contrast, BR and soil samples more frequently failed purity benchmarks, likely due to lower nucleic acid concentrations and co‐extracted organic or humic substances.

All qPCR/RTqPCR assays demonstrated robust performance, with linear standard curves spanning 5 × 10⁶ to 5 GC/reaction, amplification efficiencies of 91.3%–100% and high linearity (R ² = 0.993–0.997), consistent with MIQE guidelines (Table 1). ⁴⁰ Matrix‐associated inhibition was minimal; only three of 81 samples (one LA, two P1) exhibited ΔCq > 2, all of which were resolved following 10‐fold dilution.

Table 1.

qPCR/RT‐qPCR performance characteristics

qPCR/RT‐qPCR assays	qPCR/RT‐qPCR performance characteristics
	Efficiency	R 2	Slope	Y‐intercept
MS2 phage	100%	0.995	−3.32	37.9
PAdV	95.5%	0.997	−3.43	39.7
PCV2	91.3%	0.993	−3.55	41.2
PRoVC	99.3%	0.997	−3.34	40.3

PAdV, porcine adenovirus; PCV2, porcine circovirus type 2; PRoVC, porcine rotavirus C.

MHV nucleic acid extraction efficiency

MHV extraction efficiencies varied across sampling dates and wastewater matrices (Figure 2; Table ST3), ranging from 2.24% to 68.5% across all analysed samples. No samples were analysed on 10/07/2025. Effluent source samples (LA, SF, BR, TW), and CWS showed greater variability and higher maximum recoveries than pond samples, with peak values observed on 31/10/2025 and 24/11/2025 (up to 68.5%). Earlier sampling events (e.g., August 2025) generally exhibited lower recoveries (<35%). Across treatment stages, extraction efficiencies in P1–P3 were more consistent, typically spanning 6.90%–54.6%, with narrower interquartile ranges and median values generally between 15%–30%. Overall, extraction efficiencies differed among wastewater types, likely reflecting differences in sample composition such as suspended solids, organic matter and other constituents that can influence viral extraction efficiencies.

Figure 2.

MHV extraction efficiency across operational areas and wastewater sample stages. Box‐and‐whisker plots illustrate the distribution of MHV extraction efficiencies (%) for samples from LA (lairage), SF (slaughter floor), BR (boning room), TW (truck wash), CWS (combined waste streams), P1 (Pond 1), P2 (Pond 2) and P3 (Pond 3). For each sample type, the box represents the interquartile range (IQR; 25th–75th percentiles), the horizontal line within the box indicates the median and the plus sign (+) denotes the mean. Whiskers extend to the observed minimum and maximum extraction efficiency values.

Concentrations of viral targets across the treatment stages

Figure 3 summarizes the detection frequency of MS2 phage and viral pathogens across operational areas and treatment stages. At source locations (LA, SF, BR and TW), MS2 phage and PAdV were quantified in most samples at LA, SF and TW, while MS2 was not detected at BR. In contrast, PCV2 and PRoVC showed greater spatial variability and higher proportions of NDs, particularly at LA and TW (PCV2) and BR (PRoVC). Across the treatment train, the frequency of ND results increased progressively from CWS to P3 for all targets. PRoVC showed the greatest decline in detectability, with all P3 samples classified as ND. These patterns indicate declining viral concentrations with treatment progression. All soil samples from the effluent irrigation area were negative for all four viruses.

Figure 3.

Proportion (%) of quantified and non‐detects (NDs) results for each virus across effluent treatment stages. Sampling points include (A) LA (Lairage), SF (slaughter floor), BR (boning room) and TW (truck wash), (B) CWS (combined waste streams, P1 (Pond 1), P2 (Pond 2) and P3 (Pond 3). Bars represent the percentage of samples within each category at each treatment stage, illustrating changes in detection frequency through the treatment process.

Concentrations (log₁₀ GC/100 mL) of viruses across operational source streams (LA, SF, BR and TW) are shown in Figure 4 and Table ST4. For MS2 phage, TW exhibited the highest mean concentration (3.99 ± 1.13 log₁₀), followed by LA (3.75 ± 0.84 log₁₀) and SF (3.73 ± 1.18 log₁₀) (Table 2). In contrast, BR remained at the SLOD (2.17 log₁₀) for all sampling occasions. For PAdV, concentrations were greatest at TW (mean 7.01 ± 0.71 log₁₀ GC/100 mL), followed closely by LA (6.17 ± 1.77). SF also showed elevated levels (mean 5.93 ± 0.69), while BR recorded the lowest concentration (3.12 ± 0.87), with several ND results. PCV2 presented a different distribution: SF had the highest mean concentration (4.67 ± 1.18), with BR also showing comparatively elevated levels (3.66 ± 0.66). TW (2.97 ± 0.91) and LA (2.64 ± 0.69) had lower means, largely influenced by frequent ND results. A similar spatial pattern occurred for PRoVC, with TW again showing the highest mean concentration (4.56 ± 1.34), intermediate levels at LA (4.17 ± 1.36) and SF (3.73 ± 1.17) and the lowest values at BR (2.59 ± 0.54), which had NDs on more than half of the sampling occasions.

Figure 4.

Concentrations (log₁₀ GC/100 mL) across treatment stages, including left‐censored non‐detect observations. Non‐detect (ND) results are shown as X and were substituted with the assumed sample limit of detection (SLOD) to account for left censoring in the dataset.

Table 2.

Summary statistics of viral concentrations (log₁₀ GC/100 mL) across operational source streams

Virus	LA		SF		BR		TW
	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI
MS2 phage	3.75 ± 0.84 (9; 1)	3.12–4.38	3.73 ± 1.18 (9; 2)	2.87–4.59	2.17 ± 0.00 (9; 9)	2.17–2.17	3.99 ± 1.13 (9; 0)	3.15–4.83
PAdV	6.17 ± 1.77 (9; 1)	4.99–7.35	5.93 ± 0.69 (9; 0)	5.38–6.48	3.12 ± 0.87 (9; 3)	2.49–3.75	7.01 ± 0.71 (9; 0)	6.45–7.57
PCV2	2.64 ± 0.69 (8; 5)	2.07–3.21	4.67 ± 1.18 (8; 0)	3.73–5.61	3.66 ± 0.66 (8; 1)	3.13–4.19	2.97 ± 0.91 (8; 4)	2.23–3.71
PRoVC	4.17 ± 1.36 (9; 1)	3.10–5.24	3.73 ± 1.17 (9; 2)	2.88–4.58	2.59 ± 0.54 (9; 5)	2.18–3.00	4.56 ± 1.34 (9; 0)	3.51–5.61

n = number of sampling dates with available data (max = 9); ND = measurements left‐censored at SLOD (2.17 log₁₀ GC/100 mL).

BR, boning room; CI, confidence interval; LA, lairage area; PAdV, porcine adenovirus; PCV2, porcine circovirus type 2; PRoVC, porcine rotavirus C; SD, standard deviation; SF, slaughter floor; TW, truck wash.

Concentrations (log₁₀ GC/100 mL) of viruses across treatment stages are shown in Figure 5 and Table ST5. Viral concentrations varied markedly across the treatment stages (CWS, P1, P2 and P3), with progressive reductions observed for most targets as wastewater advanced through the system (Table 3). MS2 phage and PAdV showed the highest concentrations in CWS, followed by stepwise decreases through P1–P3, while PCV2 and PRoVC displayed greater variability and higher frequencies of NDs at later stages. For MS2 phage, mean concentrations ranged from 3.74 ± 0.75 log₁₀ GC/100 mL in CWS to 2.75 ± 0.52 log₁₀ GC/100 mL in P3, corresponding to an overall decline of nearly ~1 log₁₀ (Table 3). Temporary increases were observed in P1 during some sampling events, reflecting short‐term fluctuations in influent loads. PAdV concentrations followed a similar pattern, decreasing from 6.02 ± 0.63 in CWS to 3.62 ± 0.81 in P3, with the greatest reductions occurring between P1 and P3. Despite this decline, PAdV was still quantifiable in most P3 samples, indicating incomplete viral removal in the final pond. PCV2 exhibited higher persistence overall. Concentrations averaged 4.58 ± 0.83 log₁₀ GC/100 mL in CWS and remained detectable in P3 (3.31 ± 0.65), suggesting partial reduction but continued presence through the treatment train. In contrast, PRoVC showed the steepest decline from 3.73 ± 0.61 in CWS to 2.28 ± 0.42 in P3, with most late‐stage samples approaching the SLOD.

Figure 5.

Concentrations (log₁₀ GC/100 mL) across treatment stages, including left‐censored non‐detect observations. Non‐detect (ND) results are shown as X and were substituted with the assumed sample limit of detection (SLOD) to account for left censoring in the dataset.

Table 3.

Summary statistics of viral concentrations (log₁₀ GC/100 mL) across treatment stages

Virus	CWS		P1		P2		P3
	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI	Mean ± SD (n; ND)	95% CI
MS2 phage	3.74 ± 0.75 (9; 0)	3.16–4.32	3.97 ± 0.57 (9; 0)	3.54–4.40	3.32 ± 0.58 (9; 1)	2.88–3.76	2.75 ± 0.52 (9; 3)	2.35–3.15
PAdV	6.02 ± 0.63 (9; 0)	5.56–6.48	4.99 ± 0.47 (9; 0)	4.62–5.36	4.49 ± 0.40 (9; 1)	4.19–4.79	3.62 ± 0.81 (9; 2)	3.03–4.21
PCV2	4.58 ± 0.83 (8; 1)	3.96–5.20	4.38 ± 0.94 (9; 1)	3.71–5.05	3.89 ± 0.68 (9; 2)	3.37–4.41	3.31 ± 0.65 (9; 3)	2.81–3.81
PRoVC	3.73 ± 0.61 (9; 0)	3.27–4.19	3.12 ± 0.74 (9; 2)	2.57–3.67	2.55 ± 0.56 (9; 3)	2.14–2.96	2.28 ± 0.42 (9; 5)	1.97–2.59

n = number of sampling dates with available data (max = 9); ND = measurements left‐censored at SLOD (2.17 log₁₀ GC/100 mL).

CI, confidence interval; CWS, combined waste streams; P1, Pond 1; P2, Pond 2; P3, Pond 3; PAdV, porcine adenovirus; PCV2, porcine circovirus type 2; PRoVC, porcine rotavirus C; SD, standard deviation.

Correlation between MS2 phage and viral pathogens

Spearman correlation (r _s) analysis was used to evaluate the relationship between MS2 phage concentrations and those of PAdV, PCV2 and PRoVC across pooled and individual sample groups (Table ST6). When all data were combined, MS2 was strongly correlated with PAdV (r _s = 0.75, P < 0.001) and PRoVC (r _s = 0.68, P < 0.001), whereas the correlation with PCV2 was weaker but statistically significant (r _s = 0.35, P = 0.02). Across individual operational source streams, MS2 and PAdV concentrations were moderately to strongly associated, with significant correlations in LA (r _s = 0.80, P < 0.001), SF (r _s = 0.65, P = 0.02), TW (r _s = 0.70, P = 0.01) and CWS (r _s = 0.68, P = 0.02). In contrast, correlations between MS2 and PCV2 were consistently weak and non‐significant (rₛ = 0.20–0.45, P > 0.10). For PRoVC, moderate positive correlations were observed in LA (rₛ = 0.70, P = 0.01), SF (rₛ = 0.55, P = 0.05) and TW (rₛ = 0.60, P = 0.03), while associations in BR, CWS and the treatment stages were weaker and not statistically significant. Within the treatment stages (P1–P3), correlations between MS2 and all three target viruses were generally low and non‐significant, with r _s values ranging from 0.25 to 0.60 and P > 0.05 for most comparisons. The only marginally significant relationship occurred for MS2 vs. PAdV in P1 (r _s = 0.60, P = 0.05).

Log₁₀ reduction values (LRVs) between CWS and P3

Across the nine sampling events with paired data (eight for PCV2), apparent reductions between CWS and P3 varied by target (Table 4). PAdV exhibited the highest average reduction (mean LRV 3.01 ± 0.83; 95% CI: 2.36–3.66), with consistently large LRVs (range 1.53–3.88), reflecting frequent P3 concentrations near the SLOD. PRoVC showed moderate apparent reductions (1.48 ± 0.63; 95% CI: 0.98–1.98; range 0.82–2.75). PCV2 reductions were more variable (1.62 ± 1.46; 95% CI: 0.36–2.88; range −0.31 to 3.60), with occasional negative LRVs indicating increases at P3 relative to CWS. Negative LRVs reflect temporal variability in measured viral genome concentrations between sampling points and do not indicate biological replication or growth, as values are derived from molecular measurements of viral nucleic acids. MS2 phage displayed the lowest mean apparent reduction (1.07 ± 0.75; 95% CI: 0.47–1.66; range −0.44 to 2.34). Peak single‐date LRVs were 3.88 for PAdV (03/10/2025), 2.75 for PRoVC (16/02/2026), 3.60 for PCV2 (30/01/2026) and 2.34 for MS2 (24/11/2025). Minimal values included −0.44 for MS2 (21/08/2025), −0.31 for PCV2 (21/08/2025), 1.53 for PAdV (21/08/2025) and 0.82 for PRoVC (21/08/2025) (Table ST5).

Table 4.

Summary statistics of log reduction values (LRVs) for viral targets between combined waste streams (CWS) and Pond 3 (P3)

Microbial targets	Number of samples analysed	Mean LRV ± SD	95% CI (low–high)	Median LRV	Min LRV	Max LRV
MS2	9	1.07 ± 0.75	0.47–1.66	1.11	−0.44	2.34
PAdV	9	3.01 ± 0.83	2.36–3.66	2.98	1.53	3.88
PCV2	8	1.62 ± 1.46	0.36–2.88	1.65	−0.31	3.60
PRoVC	9	1.48 ± 0.63	0.98–1.98	1.14	0.82	2.75

CI, confidence interval; LRV, log reduction value; PAdV, porcine adenovirus; PCV2, porcine circovirus type 2; PRoVC, porcine rotavirus C.

Viable qPCR/RT‐qPCR to detect viable fraction of microbial targets in combined waste streams (CWS), pond 1 (P1), pond 2 (P2) and pond 3 (P3)

Concentrations of all viral targets declined progressively from the CWS through the pond treatment system (P1–P3), with the strongest reduction observed in P3, where most v‐PCR measurements were at the sample limit of detection (SLOD; 2.17 log₁₀ GC/100 mL) (Table 5). Across all sampling events and stages, v‐qPCR/RT‐qPCR consistently yielded lower concentrations than conventional qPCR/RT‐qPCR, reflecting the removal of signals associated with virions possessing compromised capsid integrity and other non‐viable particles. At the CWS, RT‐qPCR concentrations of MS2 phage were consistently higher than v‐RT‐qPCR concentrations, with reductions of approximately 0.60–2.60 log₁₀ GC/100 mL, whereas PRoVC showed only minor decreases (mean ~0.20 log₁₀ GC). In contrast, PAdV and PCV2 generally showed smaller differences between qPCR and v‐qPCR (≤0.30–0.50 log₁₀ for most dates), suggesting that a relatively large fraction of detected genomes remained PMAxx™‐resistant (virion‐intact) at CWS, although occasional larger deviations were observed for PCV2. On 27/02/2026, v‐qPCR/RT‐qPCR values for PAdV and PRoVC were nearly identical to qPCR/RT‐qPCR, indicating minimal loss of capsid integrity. In P1, v‐RT‐qPCR values for MS2 were uniformly at the SLOD (2.17 log₁₀ GC/100 mL), whereas RT‐qPCR concentrations ranged from approximately 3.62–4.32 log₁₀, indicating removal of viable MS2 signal during initial treatment. For PAdV and PCV2, the difference between qPCR and v‐qPCR remained modest (<0.3 log₁₀) or occasionally reversed, suggesting greater persistence of intact particles. PRoVC was undetectable by both RT‐qPCR and v‐RT‐qPCR on all dates, indicating a strong reduction by P1.

Table 5.

Mean log₁₀ gene copies (GC/100 mL) of viruses in combined waste streams (CWS), Pond 1 (P1), Pond 2 (P2) and Pond 3 (P3) determined by qPCR/RT‐qPCR and viability‐qPCR/RT‐qPCR (v‐PCR/v‐RT‐qPCR)

Sampling stages	Sampling dates	Mean log10 GC/100 mL for each target
		MS2		PAdV		PCV2		PRoVC
		RT‐qPCR	v‐RT‐qPCR	qPCR	v‐qPCR	RT‐qPCR	v‐qPCR	qPCR	v‐qPCR
CWS	30/01/2026	4.20 ± 0.03	3.60 ± 0.08	6.04 ± 0.02	5.67 ± 0.02	3.25 ± 0.40	3.63 ± 0.12	3.42 ± 0.13	3.23 ± 0.16
	16/02/2026	4.77 ± 0.01	2.17	6.53 ± 0.05	6.58 ± 0.05	5.36 ± 0.06	5.05 ± 0.02	4.05 ± 0.07	3.72 ± 0.18
	27/02/2026	3.47 ± 0.15	2.17	6.75 ± 0.05	6.76 ± 0.02	3.57 ± 0.36	4.41 ± 0.03	4.39 ± 0.08	4.35 ± 0.04
P1	30/01/2026	4.32 ± 0.05	2.17	4.54 ± 0.02	4.22 ± 0.02	4.81 ± 0.03	4.78 ± 0.01	2.17	2.17
	16/02/2026	4.08 ± 0.02	2.17	4.40 ± 0.24	3.94 ± 0.22	4.28 ± 0.01	4.40 ± 0.07	2.17	2.17
	27/02/2026	3.62 ± 0.10	2.17	4.19 ± 0.18	4.27 ± 0.23	4.46 ± 0.04	4.47 ± 0.02	2.17	2.17
P2	30/01/2026	3.81 ± 0.03	2.17	2.17	2.17	2.17	2.17	2.17	2.17
	16/02/2026	3.30 ± 0.07	2.17	4.20 ± 0.18	2.17	2.17	2.17	2.17	2.17
	27/02/2026	3.42 ± 0.10	2.17	4.59 ± 0.05	2.17	4.00 ± 0.07	4.01 ± 0.02	2.17	2.17
P3	30/01/2026	2.17	2.17	2.17	2.17	2.17	2.17	2.17	2.17
	16/02/2026	2.60 ± 0.09	2.17	3.82 ± 0.20	2.17	2.17	2.17	2.17	2.17
	27/02/2026	2.17	2.17	3.94 ± 0.34	2.17	2.17	2.17	2.17	2.17

In P2, viability signals remained at the SLOD for all targets except for a few qPCR detections of PAdV and PCV2 (e.g., 16/02/2026 and 27/02/2026). The sharp divergence between qPCR (detectable) and v‐qPCR (SLOD) for these viruses suggests that although genomic fragments persisted in the aerobic pond, the proportion of intact particles was negligible. By P3, nearly all v‐qPCR measurements were at the SLOD (2.17 log₁₀ GC), indicating the near‐complete loss of intact viral particles for all four targets at the final treatment stage. qPCR signals for MS2, PAdV and PCV2 remained detectable in some samples (e.g., MS2 at 2.60 log₁₀; PAdV at 3.82–3.94 log₁₀), reflecting residual genome fragments that remained amplifiable but likely non‐infectious. PRoVC remained undetectable by qPCR and v‐qPCR at P3 in all cases.

Clear stage‐level reductions were observed when comparing CWS to final effluent (P3). Based on qPCR signals, approximate apparent LRVs ranged from 1.7–2.6 log₁₀ reduction for MS2; 2.2–4.6 log₁₀ for PAdV; up to 3.5 log₁₀ for PCV2 and 2.5–4.1 log₁₀ for PRoVC. Using v‐PCR, all four targets exhibited maximum possible LRVs (CWS concentration −2.17), often exceeding the reductions estimated by qPCR. This pattern reflects the fact that most P3 samples were viability‐undetectable, whereas qPCR retained detectable genome fragments. Across all sampling events, conventional qPCR consistently overestimated viral persistence compared with v‐PCR. MS2 displayed the largest qPCR–v‐PCR discrepancy, especially at P1 and P2. PAdV and PCV2 exhibited smaller qPCR–v‐qPCR gaps in CWS and P1, suggesting more robust particles or protection by solids. PRoVC was rapidly destroyed, with nearly all viability signals lost after P1. The marked divergence between qPCR and v‐PCR results—particularly in P1–P3—demonstrates that substantial viral genome material persists beyond the point where structurally intact (potentially infectious) particles are no longer detected. This indicates that qPCR‐based LRVs are conservative estimates of true viral inactivation, and viability correction provides a more realistic assessment of treatment performance.

Discussion

Spatial distribution of viral loads across operational areas

Detection of all four viral targets in LA and SF samples indicates that primary contamination originating from gastrointestinal and respiratory shedding by pigs occurs during holding and slaughter. F‐RNA coliphages are normal gut flora excreted in the faeces of pigs and other livestock. ⁴¹ , ⁴² PAdV is shed in both faeces and respiratory secretions, while PRoVC is predominantly faecal in origin. ⁴³ , ⁴⁴ , ⁴⁵ PCV2 exhibits multi‐system shedding through faeces, urine, nasal secretions and saliva, resulting in widespread environmental contamination. ⁴⁴ , ⁴⁶

BR samples consistently showed lower viral concentrations and higher ND frequencies than LA and SF. This reduction reflects the progressive removal of virus‐bearing materials—faeces, urine, blood, oral and respiratory secretions—during upstream processing steps including scalding, dehairing, evisceration and carcase washing. By the boning stage, carcases have undergone substantial cleaning, and the environment is subject to stricter hygiene controls. ⁴⁷

TW exhibited the highest viral loads for MS2, PAdV and PRoVC, consistent with its role as a reservoir for accumulated faecal and other body secretion contamination from animal transport. Unlike carcases that undergo progressive decontamination, transport vehicles typically receive no cleaning until reaching the wash facility. Faecal and other residues accumulate on vehicle surfaces, and the relatively limited dilution during washing produces concentrated viral signals. Repeated washing of multiple trucks over short intervals amplifies viral accumulation as contamination from successive vehicles enters the same drainage system. This operational heterogeneity has monitoring implications. Grab samples from single locations may not fully capture variability in viral inputs across the facility. In this context, targeted sampling at high‐input nodes such as TW and CWS would provide more representative information on pathogen loads entering the system, while P3 sampling remains most relevant for assessing final effluent quality and treatment performance.

Viral reduction through the treatment stage

Viral concentrations decreased progressively from CWS through P1–P3, reflecting a combination of physical and biological processes including adsorption of viruses to suspended solids followed by sedimentation, increased residence time allowing greater exposure to inactivation processes and natural decay under environmental conditions within the pond system. PAdV exhibited the highest concentrations across all stages, consistent with its known environmental persistence and high shedding rates. PCV2 and PRoVC showed greater variability, with NDs increasing in later ponds, consistent with progressive reduction in detectable viral genomes through sedimentation, degradation and extended residence time within the treatment system. MS2 demonstrated similar declining trends, supporting the interpretation that viral reduction resulted from treatment processes rather than assay‐specific artefacts.

P3 showed the most consistent pattern of ND observations, suggesting that viral levels were approaching or falling below the sample limit of detection for multiple targets by the final treatment stage. However, intermittent detections across all targets, even at later stages, highlight inherent variability in wastewater systems influenced by hydraulic surges, loading fluctuations and upstream shedding dynamics.

MS2 phage as a process indicator

The utility of bacteriophages as surrogates for pathogen behaviour is a key question in wastewater monitoring. Strong correlations between MS2 and PAdV (rₛ = 0.75) and between MS2 and PRoVC (rₛ = 0.68) across pooled samples indicate that MS2 captures general trends for enteric viral presence. However, its weak association with PCV2 (rₛ = 0.35) and inconsistent relationships within individual treatment stages suggest that MS2 does not universally mirror all livestock virus dynamics.

This divergence is biologically plausible. MS2 is a small, non‐enveloped RNA bacteriophage infecting E. coli, with persistence dependent partly on host availability and environmental redox conditions. Adenoviruses (e.g., PAdV) and circoviruses (PCV2) are stable, host‐specific DNA viruses that interact differently with solids, organics and ammonia. Structural distinctions in capsid composition, isoelectric point and hydrophobicity translate into contrasting adsorption ⁴⁸ and photolytic outcomes. Consequently, while MS2 serves as a widely used conservative surrogate for enteric virus behaviour in wastewater studies, its performance as a universal process indicator in complex abattoir wastewater systems may be limited by virus‐specific interactions with environmental matrices. A multi‐indicator approach combining MS2 with a representative mammalian virus such as PAdV would therefore provide more robust validation of pathogen removal performance.

Log₁₀ reduction values and treatment performance

Measured LRVs between CWS and P3 ranged from approximately 1 log₁₀ for MS2 to greater than 3 log₁₀ for PAdV, with intermediate reductions for PCV2 and PRoVC. These magnitudes align with reported virus reduction in pond and lagoon systems lacking tertiary treatment. ²⁰ , ²⁶ However, substantial within‐target variability was evident. For PAdV, while mean removal exceeded 3 log₁₀, individual LRVs fluctuated from 1.5 to 3.9 across sampling dates, illustrating how short‐term changes in retention time, solids load, or sunlight exposure modulate virus fate. This variability reinforces the need to frame treatment efficiency in statistical rather than absolute terms, acknowledging that pond processes are inherently stochastic.

The persistence of quantifiable GC in P3 does not necessarily imply an environmental hazard. Genomic residues may correspond to non‐infective particles that inflate perceived exposure risk when monitoring relies solely on qPCR. This observation parallels findings from municipal wastewater studies showing that a large proportion of detectable viral genomes represent inactivated particles. ⁴⁹

Viability‐corrected assessment of viral targets

By examining both total (qPCR/RT‐qPCR) and viability‐adjusted (v‐qPCR/v‐RT‐qPCR) signals, the study moves beyond conventional LRVs to address the more nuanced issue of intact viral particles in effluent pond systems. ⁵⁰ , ⁵¹ Combining PMAxx™‐qPCR with conventional qPCR enabled the distinction between viable and non‐viable viral particles. Across all stages, viability‐corrected concentrations were 0.3–2 log₁₀ lower than total GC numbers. In P1 and especially P2, nearly all v‐qPCR results reached the sample limit of detection, indicating extensive loss of intact viral particles before final discharge. This pattern reflects the action of free radical oxidation, sunlight and microbial enzymatic activity in facultative ponds, which leads to rapid capsid degradation even when viral nucleic acid fragments remain amplifiable. ²² , ²⁴

The discrepancy between qPCR and v‐qPCR results provides an operational surrogate for inactivation dynamics. For MS2, the absence of viability signal beyond P1 suggests that most measurable genomes corresponded to non‐intact virions. PAdV and PCV2 demonstrated stronger resistance, with smaller gaps between total and viability‐corrected concentrations in early treatment stages, reflecting the stability of non‐enveloped DNA viruses. PRoVC was rapidly removed, with nearly all viability signals lost after P1.

The marked divergence between methods, particularly in P1–P3, demonstrates that substantial viral genome material persists beyond the point where intact particles remain detectable. This confirms that qPCR‐based LRVs represent conservative estimates of true viral inactivation, while viability correction provides a more realistic assessment of treatment performance. However, PMA‐based methods cannot differentiate between truly infectious and non‐infectious yet intact virions. ⁵² As a result, PMAxx™‐qPCR provides an upper bound on viable concentration and, correspondingly, a lower bound on true viral inactivation.

Regulatory implications

These findings have several implications for livestock wastewater monitoring and risk assessment. First, the presence of viral genomes in treated effluent does not necessarily equate to infection risk. Integrating viability‐corrected metrics alongside conventional qPCR/RT‐qPCR would align surveillance outcomes more closely with animal and human health risk. While acknowledging that practical constraints often necessitate conservative, genome‐based regulatory approaches, this work highlights that reliance solely on genome‐based concentration may mischaracterize system performance unless assay limitations and method complementarity are explicitly recognized.

Second, the correlation analysis demonstrates that no single indicator adequately represents all virus classes or influent streams. Validation frameworks based on multi‐target panels—combining bacteriophages with representative livestock viruses such as adenovirus—and strategic sampling at high‐input nodes would more robustly capture treatment performance and provide a stronger evidence base for assigning the extent of virus removal achieved during wastewater treatment. The consistently low viral loads observed in BR wastewater further support effective hygiene practices in the processing area and separation of this wastewater stream from higher‐contaminated wastewater streams within the study abattoir. Finally, the pronounced variability in influent viral loads highlights the importance of targeted source control. Strategic management of high‐load inputs, particularly livestock TW effluents, for example, through decontamination protocols, offers a practical approach to reduce overall system loading, improve treatment robustness and support regulatory compliance.

Overall, the study demonstrated that appropriately designed and operated pond‐based wastewater treatment systems can provide effective viral risk reduction in effluent from a swine abattoir. The reduction observed, combined with evidence that most residual viral genomes at the end of treatment are non‐intact (non‐infectious), supports regulatory confidence in the treatment performance observed in the system. The findings highlight the value of monitoring approaches that incorporate targeted source and treatment assessment and viability‐corrected interpretation of viral reduction rather than reduction estimates based on conventional genome detection methods alone.

Limitations and future directions

The use of conservative substitution for NDs likely underestimates true LRVs; actual reductions may exceed reported values. Temporal variability driven by fluctuating slaughter schedules, temperature and rainfall infiltration adds further uncertainty. Despite these sources of variability, the directional consistency of results across sampling dates implies robust system behaviour.

Future work should address standardisation of viability‐PCR protocols across virus families to improve comparability. Pairing molecular assays with cell culture infectivity or animal model studies could validate viability thresholds against true infectivity. Investigation of particle–solid partitioning and recovery biases would elucidate whether modest MHV extraction efficiencies translate into similar biases for target viruses.

Given the elevated viral loads observed in specific influent streams, particularly livestock truck wash effluents, targeted investigations of higher‐risk inputs represent an important future direction. This could include evaluation of additional upstream mitigation measures such as routine use of disinfectants during vehicle cleaning or assessment of system performance when such streams are separately managed or pre‐treated prior to mixing with lower‐risk wastewater. Conversely, worst‐case scenario analyses examining the direct land application of high viral load influent, as may occur under some operational or emergency conditions, would better define exposure risks and inform management controls. Such studies would strengthen risk‐based decision‐making by distinguishing whole‐of‐system performance from source‐specific hazards.

Conclusions

• The pond‐based swine abattoir wastewater treatment system achieved measurable reductions for all target viruses, with mean log reduction values ranging from ~1 (log₁₀ GC/100mL) for MS2 to ~3 for PAdV.

• Viability‐corrected data revealed extensive inactivation; nearly all v‐qPCR/RT‐qPCR signals were below the sample limit of detection at the final effluent stage, indicating that apparent persistence by qPCR/RT‐qPCR represented non‐intact genomes.

• MS2 phage correlated strongly with enteric PAdV and PRoVC but poorly with PCV2, suggesting it may be a suitable indicator for enteric‐virus behaviour but not for DNA viruses with distinct stability profiles.

• Integration of viability‐qPCR/RT‐qPCR improves the precision of risk interpretation by distinguishing potentially infectious from non‐infectious viral particles.

• Overall, the system acheived substantial multi‐stage viral reduction, and the low level detection of intact viral particles in the final effluent suggests a reduced viral risks for irrigation reuse under normal operating conditions

• Future research should emphasize long‐term monitoring, solid–liquid partitioning assessments and optimisation of hydraulic retention or disinfection stages to further enhance viral load reduction in livestock processing wastewater.

Conflicts of interest and sources of funding

The authors declare no conflicts of interest or sources of funding for the work presented here.

Supporting information

Table ST1. qPCR/RT‐qPCR primer/probe sets, concentrations and thermal cycling conditions used in this study.

Table ST2. Concentration (ng/μL) and spectrophotometric purity (A _260/280 and A _260/230 ratios) of extracted nucleic acid samples from wastewater samples across operational sources and treatment stages.

Table ST3. MHV extraction recovery efficiency across operational source streams and wastewater sample stages.

Table ST4. Concentrations (log₁₀ GC/100 mL) of viruses across operational source streams (LA, SF, BR and TW), including left‐censored non‐detect (ND) data. Bold‐faced values represent ND results, which were substituted with the assumed sample limit of detection (SLOD).

Table ST5. Concentrations (log₁₀ GC/100 mL) of viruses across treatment stages, including left‐censored non‐detect data. Bold‐faced values represent non‐detect (ND) results, which were substituted with the assumed sample limit of detection (SLOD).

Table ST6. Nonparametric Spearman correlation (r _s) analysis between MS2 phage concentrations and concentrations of other viruses in pooled operational source streams and across treatment stages.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.