Multi-Isotope Signatures in Modern Pig Bones: Evaluating Dietary and Hydrological Controls on Provenance in South Korea

Abstract

Multi-isotope analysis of animal tissues is a powerful tool for provenance studies; however, its application to modern livestock is complicated by the widespread use of manufactured feed and managed water sources. In this study, we investigated the mechanisms by which dietary and hydrological isotope signals are incorporated into modern pig bone tissues and evaluated whether bone collagen (BC) and phosphate (BP) preserve interpretable provenance information. ⁸⁷Sr/⁸⁶Sr, δ¹³C, δ¹⁵N, and δ¹⁸O values were measured in pig bone, BC, and BP collected from 17 livestock farms across South Korea. Bone isotope compositions were compared with corresponding feed and groundwater values to assess the transfer of dietary and hydrological signals into skeletal tissues. Bone ⁸⁷Sr/⁸⁶Sr ratios ranged from 0.70991 to 0.72280 and largely overlapped with feed values (0.71009–0.71226), whereas groundwater values differed significantly from bone values (p < 0.0001), indicating that Sr incorporation was more strongly influenced by diet than by drinking water. However, the relative contributions of feed, groundwater and other sources were not quantified for source apportionment. δ¹³C_BC values ranged from −17.6‰ to −13.4‰, and δ¹⁵N_BC values ranged from 0.74‰ to 5.31‰. Both isotopic systems were significantly enriched relative to feed (p < 0.0001), with mean offsets of 1.69‰ and 1.72‰ for δ¹³C and δ¹⁵N, respectively. Both δ¹⁸O_BC and δ¹⁸O_BP were positively correlated with δ¹⁸O_Water, although BP exhibited relatively stronger explanatory power (R ² = 0.61) than BC (R ² = 0.40). These results suggest that modern pig bones retain measurable dietary and hydrological isotope signals; however, the use of manufactured feed can possibly weaken or obscure local geological provenance signatures. Notably this study provides a multi-isotope baseline for pig farms in South Korea and highlights both the potential and limitations of skeletal isotope analysis for provenance investigations of modern livestock systems.

1. Introduction

Geographical provenance tracing of plants, animals, and humans is important in ecology, environmental science, forensics, geology and archeology. , Over the past two decades, stable isotope analysis has become a widely applied approach for investigating biological, geological, and ecological processes, because isotopic compositions reflect environmental heterogeneity and biogeochemical pathways. − During transfer from the environment into biological systems, isotopic signatures derived from local geology, climate, water, and diet are incorporated into plant and animal tissues, thereby preserving information on origin and ecological interactions. ,

Among the isotopic systems used in provenance research, strontium (Sr) isotope ratios (⁸⁷Sr/⁸⁶Sr) are particularly valuable because they reflect regional geological characteristics. , As bedrock undergoes weathering, Sr is released into soils and waters and subsequently incorporated into plants and animals via diet and drinking water, with minimal biological fractionation. Consequently, ⁸⁷Sr/⁸⁶Sr values in biological tissues can closely reflect environmental signatures and serve as robust indicators of geographic origin.

Stable carbon (δ¹³C) and nitrogen (δ¹⁵N) isotopes provide complementary information on dietary composition and trophic relationships. Carbon isotope ratios primarily reflect consumption behavior, distinguishing between C3 and C4 diets, whereas nitrogen isotope ratios are associated with trophic positions and dietary protein sources. ,, Together, these isotopic systems provide insight into feeding behavior and dietary composition. Hydrogen (δ²H) and oxygen (δ¹⁸O) isotope ratios in animal tissues are strongly influenced by ingested water. , Because meteoric water isotopes vary systematically with latitude, altitude, and climate, tissue δ²H and δ¹⁸O values can preserve hydrological information relevant to provenance.

In recent years, stable isotope analysis of tissues such as hair, feathers, nails, bones, and tooth enamel has been increasingly used to study migration, diet, habitat use, and food provenance. , Among these tissues, bone collagen (BC) and phosphate (BP) are particularly valuable because they preserve isotopic information over extended periods and are relatively resistant to degradation. BC primarily reflects dietary protein, whereas hydroxyapatite oxygen isotopes are more closely linked to body water derived from drinking water and food. , These properties have led to their extensive application in ecological and provenance-based studies.

To date, only a few studies have evaluated the applicability of skeletal isotope analysis to modern livestock production systems. In contemporary animal agriculture, globalization and the widespread distribution of commercial feed can obscure local environmental isotope signals. , Manufactured feed and managed water supplies can weaken the relationship between animal tissues and local geological or hydrological conditions, thereby limiting the effectiveness of isotope-based provenance approaches. , Therefore, isotopic signatures preserved in modern livestock tissues may reflect complex mixtures of regional and nonlocal inputs rather than simple environmental baselines.

Considering current challenges, determining the mechanisms by which isotopic signals are incorporated into modern animal tissues is essential for improving provenance and traceability frameworks. Understanding the relative impacts of diet, water, and environmental sources on tissue isotope compositions is particularly important for livestock raised in intensive production systems, where feed ingredients and water sources may be spatially decoupled from local environmental conditions.

To address the above-mentioned knowledge gap, we investigated the multi-isotope compositions of bone, BC and BP in modern pigs collected across South Korea. We analyzed their ⁸⁷Sr/⁸⁶Sr, δ¹³C, δ¹⁵N, and δ¹⁸O values and compared them with corresponding feed and water isotope compositions. This study evaluated the extent to which modern pig bones retain interpretable provenance information and established the first skeletal multi-isotope baseline for pigs in South Korea.

2. Materials and Methods

2.1. Sample Collection

A total of 50 modern pig bone samples were collected from 17 pig farms across South Korea (Figure ). In commercial swine production systems, low-weight individuals, commonly referred as runt pigs, are frequently culled to optimize productivity. Accordingly, these individual pigs (between 50 and 100 days of age) were used as study material. Two to three pig bone samples were collected from each farm. To minimize variability associated with production systems, all sampled farms raised pigs under strict standards (e.g., breeding stock, artificial insemination, feed management) of a Pig Farmers Cooperative. The pigs represented Yorkshire–Landrace–Duroc crossbreeds. Following collection, bone samples were placed in clean, sealed polyethylene zip-lock bags and transported to the laboratory in an insulated ice box containing ice packs. Upon arrival at the laboratory, the samples stored at −20 °C until pretreatment. Feed samples (feed for gestating sow, lactating sow, and piglet) were obtained directly from the feed supplier during the same period as bone sampling. As the pigs were raised under a strictly controlled production system in this study, all sampled farms utilized the same feed.

1.

A map showing sampling locations.

Pigs were not sacrificed specifically for this study, as all samples were collected after legally authorized slaughter. Therefore, in accordance with Korean regulations, ethical approval was not required, and the study was exempted from ethical review.

2.2. Sample Preparation

Soft tissues adhering to bone surfaces were removed using a ceramic knife. The whole-bone samples were frozen in liquid nitrogen and fragmented using a hammer. The bone fragments (approximately 1–3 mm) were hand-picked. Samples were subsequently rinsed with deionized (DI) water in an ultrasonic bath (JAC-2010, Kodo, Korea) for 1 h to remove any residue. Following cleaning, bones were defatted in 8 mL of a 2:1 (v/v) dichloromethane:methanol solution in an ultrasonic bath for 30 min. The solvent mixture was prepared using HPLC-grade dichloromethane and methanol (JT Baker). After ultrasonication, the solvent was discarded and replaced with fresh solvent mixture, followed by an additional 30 min of ultrasonication. This procedure was repeated until the solvent remained clear. Samples were then rinsed sequentially with acetone (JT Baker) and DI water. Cleaned bone samples were lyophilized (FDU-2100, EYELA, Japan) for 3 days prior to further analytical processing.

The collected feed samples were freeze-dried for 3 days and subsequently ground into a fine homogeneous powder using a ball mill (MM-400, Retsch, Germany) prior to isotope analysis.

2.2.1. Collagen Extraction for Carbon, Nitrogen, and Oxygen Isotope Analysis

Collagen extraction for carbon and nitrogen isotope analysis was conducted following Sykut et al. Approximately 300 mg of pretreated bone sample was demineralized in 6 mL of 0.5 M HCl (ACS grade; Sigma-Aldrich) for 24 h, with the acid solution replaced after approximately 16 h. Following demineralization, the samples were washed repeatedly with DI water. During each washing cycle, the samples were briefly immersed in DI water and gently agitated (5 min), after which the wash solution was discarded and replaced with fresh DI water for the next washing cycle. Washing was considered complete when the final wash solution was visually clear and achieved a neutral pH. The pH of the wash solution was tested using pH indicator strips (range 0–14, Macherey-Nagel, Duren, Germany). The demineralized and DI water-washed samples were then subjected to gelatinization.

For gelatinization, demineralized residues were transferred to a 50 mL centrifuge tube and treated with 5 mL of 0.001 M HCl. Gelatinization was performed by incubating the samples overnight at 60 °C. Following incubation, the samples were centrifuged at 3000 rpm for 10 min (Combi 408, Hanil Science, Korea). The supernatant was carefully collected into clean tubes without disturbing the remaining insoluble material and lyophilized for 3 days to obtain purified collagen.

Collagen extraction for δ¹⁸O analysis was performed using a modified protocol based on the methods of von Holstein et al. and Tuross et al. Pretreated bone samples were demineralized in 50 mL of 0.5 M EDTA (pH 7.5; Sigma-Aldrich) at room temperature for 4 weeks until the samples became pale yellow to translucent, indicating complete demineralization. The EDTA solution was replaced every 3–4 days throughout the demineralization process. Following demineralization, samples were rinsed seven times with DI water.

To remove humic substances, samples were treated with 0.1 M NaOH (Fisher Chemical) and subsequently rinsed seven times with DI water. Cleaned samples were then lyophilized for 3 days prior to isotopic analysis. Although oxygen exchange between collagen and extraction reagents has been reported during pretreatment, the magnitude of preparation-induced exchange was not independently evaluated in this study through a pretreatment-control or re-extraction experiment.

2.2.2. Phosphate Preparation

Phosphate was extracted following Bera et al. Defatted bone residues remaining after collagen extraction were ground into a fine powder using a ball mill (MM-400; Retsch, Germany). Approximately 40 mg of powdered bone was treated with 2 mL of 2.5% sodium hypochlorite (NaOCl; SamChun, Korea) for 24 h to remove residual organic matter. Samples were centrifuged at 3000 rpm for 10 min. Supernatants were discarded and precipitates were subsequently washed with DI water to remove residual NaOCl. The precipitates were then treated with 2 mL of 0.125 M NaOH (Fisher Chemical) for 48 h on an orbital shaker (SHO-2D, Daihan Scientific, Korea), followed by centrifugation at 3000 rpm for 10 min. After removal of the supernatant, residues were washed five additional times with DI water. Afterward, 2 N HF (Sigma-Aldrich) was added to the residues and left undisturbed for 24 h to completely precipitate CaF₂. Following this, centrifugation was done and the supernatants were collected and neutralized with 3 mL of 2 M KOH (Sigma-Aldrich). Each solution was adjusted to a final volume of 200 mL using DI water. Silver phosphate (Ag₃PO₄) was precipitated by adding 30 mL of silver nitrate solution containing 0.2 M AgNO₃, 0.35 M NH₄NO₃, and 0.75 M NH₄OH, followed by heating at 70 °C for 3 h. The resulting precipitate was filtered through a 0.2 μm cellulose acetate membrane filter (Advantec, Japan) and dried at 50 °C for isotopic analysis.

2.3. Isotope Analysis

2.3.1. Strontium Isotope Analysis

Strontium concentrations and ⁸⁷Sr/⁸⁶Sr ratios were determined following modified protocols from Knudson and Price. Bone surfaces were mechanically cleaned using a ceramic knife, and trabecular tissue was removed with a dental drill. Samples were subsequently washed with DI water and 30% hydrogen peroxide (H₂O₂) in an ultrasonic bath to remove residual organic material, followed by ashing in furnace (FX-14, Daihan Scientific, Korea) at 750 °C for 8 h.

For Sr concentration analysis, approximately 5 mg of ashed bone powder was dissolved in 500 μL of 5 M HNO₃, evaporated at 120 ∼ 160 °C, and subsequently redissolved in 5% HNO₃. Strontium concentrations were measured using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Optima 8300, PerkinElmer, USA).

For ⁸⁷Sr/⁸⁶Sr analysis, samples were redissolved in 8 M HNO₃, and Sr was isolated using ion-exchange chromatography (Eichrom Industries, USA). Isotope ratios were measured using multicollector inductively coupled plasma mass spectrometry (MC-ICP-MS; Neptune, Thermo Fisher Scientific, Germany). The ⁸⁷Sr/⁸⁶Sr ratios were normalized using bracketing method with NBS987 as the external reference material. The average ⁸⁷Sr/⁸⁶Sr ratio obtained for NBS987 was 0.710246 ± 0.000020 (2SD, n = 17).

2.3.2. Carbon (δ¹³C) and Nitrogen (δ¹⁵N) Isotope Analysis of BC

Approximately 0.05 mg of extracted and purified BC was weighed into tin capsules (3.5 × 1.5 mm; Elemental Microanalysis, UK). δ¹³C and δ¹⁵N values were measured using an isotope ratio mass spectrometer (IRMS; VisION, Isoprime, UK) coupled to an elemental analyzer (Elementar, Langenselbold, Germany).

Measured δ¹³C and δ¹⁵N values were calibrated using USGS reference materials (USGS-40, USGS-43, USGS-88, and USGS-89). All samples and reference materials were measured in duplicate. Analytical precision was ± 0.1‰ for δ¹³C and ± 0.15‰ for δ¹⁵N.

2.3.3. Oxygen (δ¹⁸O) Isotope Analysis of BC and BP

For oxygen isotope (δ¹⁸O) analysis, lyophilized collagen and phosphate samples (0.15–0.20 mg) were weighed into silver capsules and pyrolyzed at 1450 °C to generate CO gas. The generated gas was separated chromatographically and δ¹⁸O measurements were done using IRMS (Delta V Advantage and Delta V Plus; Thermo Fisher Scientific, Bremen, Germany) coupled with thermal conversion elemental analyzers (TC/EA; Flash 2000HT, Thermo Fisher Scientific, Bremen, Germany).

δ¹⁸O_BC values were normalized using USGS reference materials (USGS-88, USGS-89, KHS, and CBS). δ¹⁸O_BP values were calibrated using Ag₃PO₄ laboratory standards (YR-1a and YR-3.2 assigned with −5.49‰ and 33.63‰, respectively.). All samples and reference materials were analyzed in triplicate.

2.3.4. Oxygen (δ¹⁸O) Isotope Ratios of Water Used in Pig Production

Groundwater samples used for pig rearing were collected from each farm at the time of bone sampling. The samples were transferred into LDPE bottles without headspace to minimize evaporation. The samples were stored at a refrigerator until analysis. Samples (0.25 mL) were transferred into 5 mL glass vials and equilibrated with CO₂ gas. Vials were sealed and incubated at 23 °C for 24 h to achieve isotopic equilibrium. The CO₂ gas was subsequently extracted and cryogenically purified. Oxygen isotope ratios were measured using a dual-inlet IRMS (Isoprime 100, UK) and calibrated against VSMOW and USGS water standards (W-67400, W-32593, and W-43156).

All isotope analyses were conducted at the Korea Basic Science Institute (KBSI), Korea. Stable isotope values are reported using delta (δ) notation:

$$\delta (\mathrm{‰})=[(R_{\mathrm{sample}}/R_{\mathrm{reference}})-1]\times 1000$$

where R represents ¹³C/¹²C, ¹⁵N/¹⁴N, or ¹⁸O/¹⁶O ratios. Carbon isotope values are reported relative to VPDB, nitrogen isotope values relative to atmospheric N₂, and oxygen isotope values relative to VSMOW. Reference materials used and standardization procedures used for the stable isotope analyses are described in the Supporting Information (Table S1, Figures S1 and S2).

2.3.5. Data Analysis

Differences in ⁸⁷Sr/⁸⁶Sr between bone and feed or groundwater, and δ¹³C and δ¹⁵N between BC and feed were evaluated using Mann–Whitney test (p ≤ 0.05). A nonparametric Mann–Whitney U test as it does not assume normality of the data. One-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test (p < 0.05) was used to compare isotopic difference between BC, BP, feed, and groundwater. Linear regression was used to examine the relationships between skeletal and groundwater isotope values and between δ¹⁸O values of BC and BP. The slopes of regressions were compared using analysis of covariance (p < 0.05). To avoid pseudoreplication, the isotope values of bone, BC, and BP from pigs sampled within the same farm were averaged to obtain mean values for statistical comparisons. All statistical analyses were conducted using XLSTAT-Pro v.13.2.1.0 (Addinsoft, New York, NY, USA).

3. Results

3.1. Sr Isotope Ratios in Bone and Feed

The ⁸⁷Sr/⁸⁶Sr ratios in bone ranged from 0.70991 to 0.72280. Approximately 85% of bone sample exhibited values <0.71279. ⁸⁷Sr/⁸⁶Sr_Feed values were narrowly distributed, with a mean of 0.71108 ± 0.0009 and range of 0.71009–0.71226 (Table ; Figure a).

1. Descriptive Statistics of ⁸⁷Sr/⁸⁶Sr, δ¹³C, δ¹⁵N, and δ¹⁸O in the Pig Bone, BC, BP, Feed, and Groundwater (Water) in South Korea .

	87Sr/86SrBone	87Sr/86SrFeed	87Sr/86SrWater	δ13CBC	δ13CFeed	δ15NBC	δ15NFeed	δ18OBP	δ18OBC	δ18OFeed	δ18OWater
Range	0.013	0.002	0.013	4.18	3.35	4.57	1.15	3.01	3.10	1.58	3.19
Mean	0.71228	0.71108	0.71466	–16.5	–18.2	3.12	1.40	13.7	7.34	24.6	–7.91
SD(n-1)	0.003	0.001	0.003	0.95	1.31	1.04	0.44	0.78	0.83	0.68	1.00
MAD	0.002	0.001	0.002	0.69	1.07	0.81	0.35	0.66	0.69	0.56	0.83
MDAD	0.0006	0.0004	0.0009	0.48	0.28	0.63	0.17	0.58	0.62	0.37	0.62

^aSD, MAD, and MDAD are standard deviation, mean absolute deviation, and median absolute deviation, respectively.

2.

(a) Box plot showing comparative distributions of ⁸⁷Sr/⁸⁶Sr isotopic signatures among pig bone, pig feed and groundwater samples. (b) and (c) are mean bar charts showing differences in the isotopic composition of pig bone with feed (p = 0.21) and groundwater (p < 0.0001), respectively, with Mann–Whitney test. Vertical lines in (b) and (c) are standard deviations. (d) shows offset value (Δ) of groundwater (Δ⁸⁷Sr/⁸⁶Sr_Water‑Bone) (open circles) and average feed (Δ⁸⁷Sr/⁸⁶Sr_Feed‑Bone) (asterisk) from bone. Plus (+) sign represents absolute deviation. (e) Linear regression biplot of ⁸⁷Sr/⁸⁶Sr for pig bone and groundwater samples. Solid line is the best fit line, and dotted lines are confidence intervals (95%).

Mann–Whitney tests indicated no significant difference between bone and feed ⁸⁷Sr/⁸⁶Sr ratios (U = 195; p = 0.21; Figure b). In contrast, a significant difference was observed between bone and groundwater values (U = 453; p < 0.0001; Figure c). The mean groundwater-to-bone offset (Δ⁸⁷Sr/⁸⁶Sr_Water–Bone) was 0.002 (Figure d), whereas the feed-to-bone offset (Δ⁸⁷Sr/⁸⁶Sr_Feed‑Bone) was – 0.001. Linear regression analysis revealed a statistically significant but weak relationship between ⁸⁷Sr/⁸⁶Sr_Bone and ⁸⁷Sr/⁸⁶Sr_Water (R ² = 0.27; p = 0.0001; Figure ).

3.

(a) Box plot showing comparative distributions of δ¹³C (‰) between pig BC and pig feed recorded from South Korea. (b) Mean bar charts showing differences in the δ¹³C isotopic composition of pig BC with feed (p < 0.0001) of Mann–Whitney test. (c) Box plot showing comparative distributions of δ¹⁵N (‰) between pig BC and pig feed. (d) Mean bar charts showing differences in the δ¹⁵N pig BC and δ¹⁵N feed (p < 0.0001) of Mann–Whitney test. (e) and (f) show offset values (Δ) of δ¹³C and δ¹⁵N, respectively. The plus sign and horizontal line represent absolute mean and median values, respectively.

3.2. Carbon and Nitrogen Isotope Ratios in BC and Feed

δ¹³C_BC values ranged from −17.6‰ to −13.4‰, with a mean of −16.5 ± 0.94‰ (Figure a). δ¹³C_Feed values ranged from −20.4‰ to −17.0‰, with a mean of −18.2 ± 1.20‰. Mann–Whitney tests indicated a significant difference between BC and feed δ¹³C values (U = 278, p < 0.0001; Figure b).

δ¹⁵N_BC values ranged from 0.74‰ to 5.31‰, averaging 3.12 ± 1.03‰, whereas δ¹⁵N_Feed values ranged from 0.63‰ to 1.78‰, averaging 1.40 ± 0.41‰ (Figure c). A significant difference was observed between δ¹⁵N_BC and δ¹⁵N_Feed values (U = 287; p < 0.0001; Figure d).

The collagen-to-feed carbon isotope offset (Δ¹³C_BC–Feed) ranged from 0.6‰ to 4.8‰ (Figure e), whereas the nitrogen isotope offset (Δ¹⁵N_BC–Feed) ranged from 0.66‰ to 3.91‰ (Figure f). Dual-isotope plots (δ¹³C vs δ¹⁵N) showed separation between BC and feed isotope compositions. Compared to previously reported Korean pork samples (δ¹³C = −19.5‰ to −16.1‰, δ¹⁵N = 2.8‰ to 4.3‰). δ¹⁵N BC samples exhibited similar δ¹⁵N ranges but nonoverlapping δ¹³C ranges (Figure ).

4.

δ¹³C and δ¹⁵N values in BC of pig, and its feed used by different stages of pigs from pig house in South Korea.

3.3. Oxygen Isotope Ratios in BC, BP, Water, and Feed

δ¹⁸O_BC ranged from 5.8‰ to 8.9‰, averaging 7.34 ± 0.82‰, whereas δ¹⁸O_BP ranged from 12.3‰ to 15.3‰, averaging 13.7 ± 0.77‰. δ¹⁸O_Feed values were substantially higher, ranging from 24.0‰ to 25.6‰ (mean = 24.6 ± 0.62‰). In contrast, δ¹⁸O_Water values were much lower, ranging from −9.26‰ to −6.07‰ (mean = −7.91 ± 0.97‰). Significant differences were observed among all sample types (p < 0.0001; Figure ).

5.

(a) Box plot showing comparative distributions of δ¹⁸O values in pig BC and BP, feed, and water from South Korea. The plus sign and horizontal line represent mean and median. (b) Mean bar charts showing Tukey’s pairwise differences in the δ¹⁸O values between four types of samples. Different lowercase letters represent results of Tukey’s multicomparison test following ANOVA showing significant differences between sample types. (c) Average offset values (Δ) of δ¹⁸O in BC, BP and feed.

Both δ¹⁸O_BC and δ¹⁸O_BP exhibited significant positive relationships with δ¹⁸O_Water (BC: R ² = 0.40 and p < 0.01; BP: R ² = 0.61 and p < 0.001; Figure ). The regression slope was slightly higher for BP (0.53) than for BC (0.48). However, the slopes were not significantly different (p = 0.74).

6.

Linear regression biplot showing best fit line of the relationship of (a) δ¹⁸O_BC–δ¹⁸O_Water and (b) δ¹⁸O_BP–δ¹⁸O_Water. (c) Biplot of δ¹⁸O_BC–δ¹⁸O_Water showing comparison of this study’s result against previous studies. (d) Linear regression biplot showing best fit line of the relationship between oxygen isotope of BP (δ¹⁸O_BP) and water (δ¹⁸O_Water) from South Korea.

A significant positive relationship was also observed between δ¹⁸O_BC and δ¹⁸O_BP (R ² = 0.36; p < 0.0001; Figure ). Compared to those for groundwater, Δ¹⁸O offsets were 14.1–16.8‰ for BC (mean = 15.5 ± 0.79‰) and 20.0–22.7‰ for BP (mean = 21.7 ± 0.64‰). Notably, BP consistently exhibited greater ¹⁸O enrichment than BC, with differences of approximately 5–7‰.

7.

(a) Biplot showing a relationship between δ¹⁸O of BC and BP (a). (b) Δ¹⁸O_BC vs Δ¹⁸O_BP with respect to water.

4. Discussion

4.1. Relationship of Bone ⁸⁷Sr/⁸⁶Sr with Water and Feed ⁸⁷Sr/⁸⁶Sr

Bone ⁸⁷Sr/⁸⁶Sr values primarily reflect dietary and drinking-water inputs, whereas geological substrate ultimately govern environmental Sr isotopic baselines. , In this study, ⁸⁷Sr/⁸⁶Sr values of bone, feed, and groundwater fell within the previously reported ranges for Korean geological and environmental materials. ,,, However, isotopic variability was relatively limited, possibly reflecting the predominance of granitic bedrock across many sampled regions, which may reduce regional environmental heterogeneity (Figure ).

The weak relationship between groundwater and bone (R ² = 0.27) suggests that groundwater has limited direct influence on bone ⁸⁷Sr/⁸⁶Sr values. Although statistically significant differences were observed between groundwater and bone ⁸⁷Sr/⁸⁶Sr values, the relatively low explanatory power suggests that groundwater alone cannot adequately account for the variation in skeletal Sr isotope compositions. These results imply the involvement of additional Sr sources or mixing processes.

The close correspondence between feed and bone ⁸⁷Sr/⁸⁶Sr values probably suggests that dietary inputs are the dominant influence on skeletal Sr incorporation. This interpretation is consistent with previous reports that diet often exerts a stronger influence than drinking water on tissue Sr isotope compositions. , Because Sr is absorbed through the gastrointestinal tract and incorporated into mineralized tissues with minimal isotopic fractionation, the isotopic composition of feed may exert a substantial influence on skeletal Sr signatures. However, the broad overlap between bone and feed ⁸⁷Sr/⁸⁶Sr values does not necessarily establish feed as the dominant Sr source. Quantitative source apportionment would require information on Sr concentrations, necessarily establish feed as the dominant Sr source. Previous controlled experiments have demonstrated that bulk-feed ⁸⁷Sr/⁸⁶Sr values may differ from those of the bioavailable fraction and that water can become an important source of skeletal Sr when its Sr concentration is sufficiently high. These parameters were not determined in the present study; therefore, a concentration-weighted mixing analysis could not be performed.

Moreover, the occurrence of bone ⁸⁷Sr/⁸⁶Sr values outside the measured feed range indicates that the sampled feed alone cannot account for the full range of skeletal Sr isotope compositions, suggesting contributions from unmeasured and/or temporally variable Sr sources. Additional exogenous factors, including incidental soil ingestion and atmospheric dust deposition, possibly contribute to the variability in bone ⁸⁷Sr/⁸⁶Sr values. The rooting behavior of pigs can result in unintentional ingestion of soil and dust particles, potentially altering internal Sr pools. − Moreover, western regions of Korea are periodically affected by long-range dust transport originating from the Loess Plateau and arid regions of China and Mongolia. If the isotopic composition of deposited dust differs from that of local groundwater or feed, these materials may indirectly contribute to variation in skeletal ⁸⁷Sr/⁸⁶Sr values.

As commercially manufactured feeds commonly contain ingredients sourced from multiple geographic regions, skeletal Sr isotope signatures may reflect mixed international inputs rather than strictly local environmental sources. , This possibility complicates geographic assignment based solely on ⁸⁷Sr/⁸⁶Sr values. The apparent decoupling between δ¹⁵N and ⁸⁷Sr/⁸⁶Sr values in this study further supports the interpretation that distinct components of the feeding system independently influence these isotope systems (Figure ).

8.

Biplot of δ¹⁵N_BC and ⁸⁷Sr/⁸⁶Sr of bone of pigs in South Korea. Black line is the best fit line, and dotted lines are upper and lower 95% confidence intervals.

4.2. Relationship of Carbon and Nitrogen Isotopes in BC with Those in Feed

In this study, δ¹³C_BC values exhibited relatively limited variability despite the broad geographic distribution of the sampled farms (Figure ), suggesting isotopically homogeneous dietary inputs. This restricted range indicated that the pigs consumed feeds with broadly similar isotopic baselines, consistent with the standardized formulation of commercial livestock feed. Because BC has a relatively slow turnover rate and integrates dietary signals over the collagen formation, the isotopic composition of BC in pigs aged 50–100 days is expected to reflect dietary inputs acquired during gestation, lactation, and postweaning stage. Therefore, feed samples corresponding to all three feeding stages were considered.

BC was enriched in ¹³C relative to feed, with a mean offset of approximately 1.7‰. Although this enrichment is lower than the 3–6‰ consumer-diet discrimination commonly reported for many vertebrates, − it indicates strong dietary correspondence and demonstrates that BC largely preserves feed-derived carbon isotope signatures under current rearing conditions. Rather than reflecting substantial isotopic modification during tissue synthesis, the relatively small Δ¹³C_BC–Feed suggests close isotopic correspondence of dietary inputs.

This comparatively limited carbon isotope offset may reflect several interacting factors, including tissue-specific metabolic routing, variation in dietary lipid content, and the turnover dynamics of BC. Compared to soft tissues such as muscle, BC exhibits slower turnover and integrates dietary inputs over longer time scales, thereby producing a temporally averaged isotopic signal.

Nitrogen isotopic compositions showed consistent enrichment in BC compared to those in feed, with a mean Δ¹⁵N_BC‑Feed of approximately 1.7‰, consistent with trophic enrichment processes (Figure f). Although this enrichment is lower than the commonly reported 3–5‰ consumer-diet offset, the direction and magnitude of enrichment remain broadly consistent with theoretical expectations for dietary nitrogen transfer. The relatively small nitrogen isotope offset may reflect both the isotopic composition of feed ingredients and the predominantly plant-based nature of swine diets.

The observed δ¹⁵N values in both feed (1.40 ± 0.41‰) and BC (3.12 ± 1.03‰) possibly reflect feed formulations dominated by legumes, particularly soybean-based ingredients. Because legumes acquire nitrogen via biological fixation of atmospheric N₂, they typically exhibit comparatively low δ¹⁵N values, and their incorporation into livestock feed lowers the overall dietary isotopic baseline. Consequently, δ¹⁵N_BC values primarily reflect this low-baseline dietary input rather than strong trophic enrichment alone. These findings highlight the importance of interpreting consumer isotope compositions within the isotopic context of feed composition.

Overall, the combined δ¹³C and δ¹⁵N results suggest that BC preserves a strong dietary signal while integrating cumulative feed inputs over extended periods. In modern production systems, skeletal isotope compositions primarily reflect homogenized commercial feed mixtures rather than fine-scale geographic variation. As industrial feeds contain multiple ingredients with potentially diverse geographic and isotopic origins, accounting for feed heterogeneity is essential when applying BC isotopes to provenance and traceability studies.

4.3. Relationship of δ¹⁸O Values in BC and BP with Those in Water and Feed

δ¹⁸O values in BC and BP primarily reflect oxygen derived from ingested water, with additional contributions from dietary oxygen and physiological processes. ,, Consistent with this, both tissues exhibited significant positive relationships with drinking-water δ¹⁸O (Figure ), indicating an association between skeletal and water oxygen isotope compositions under the condition examined. This finding is consistent with previous experimental and observational reports that tissue δ¹⁸O values track environmental water inputs. ,

The regression slopes for BC (0.48) and BP (0.53) indicate that variations in skeletal δ¹⁸O values responds proportionally to changes in drinking-water isotopic composition. Although the slopes were statistically indistinguishable, BP exhibited relatively greater explanatory power (R ² = 0.61) than BC (R ² = 0.40), indicating a stronger association with drinking-water δ¹⁸O. This difference possibly reflects tissue-specific pathways of oxygen incorporation. BP is formed via mineralization processes closely linked to body water, whereas BC incorporates oxygen through additional metabolic and biochemical pathways, resulting in greater isotopic complexity. The consistent enrichment of BP relative to BC by approximately 5–7‰ reflects systematic fractionation differences associated with tissue synthesis and bone remodeling. Despite this offset, the significant positive relationship between BC and BP indicates the coordinated incorporation of oxygen from shared physiological reservoirs.

Although drinking water accounted for a substantial proportion of variation in skeletal δ¹⁸O values, residual variability likely reflects contributions from dietary oxygen and physiological factors. δ¹⁸O_Feed values were strongly enriched relative to groundwater δ¹⁸O values but exhibited limited variability among farms, consistent with the use of standardized commercial feed formulations. Consequently, dietary oxygen likely contributed to overall enrichment while exerting relatively little influence on between-farm variation.

Overall, these results suggest that both BC and BP reliably preserve hydrological isotope information while simultaneously integrating dietary and physiological influences. Although this difference in R ² indicates a closer statistical fit for BP but does not, by itself, demonstrate its greater sensitivity to drinking-water δ¹⁸O. Both tissues therefore show potential for hydrological provenance assessment. However, provenance assessment is limited by the temporally limited collection groundwater sample from each farm. Because BC and BP integrate oxygen inputs over extended periods, a single water sample may not fully represent temporal variation in the water consumed during tissue formation. Repeated seasonal sampling can help determine the sensitivity of BP and BC to hydrological isotope variation. Additionally, von Holstein et al. demonstrated that although exchange during EDTA demineralization at neutral pH 7.5 was substantially lower, collagen oxygen can exchange with reagent water and atmospheric moisture. Preparation-induced oxygen exchange in BC was not independently quantified in this study. The observed relationship between BC δ¹⁸O and groundwater therefore represents an association under the standardized preparation conditions used in this study and some exchange with reagent water or atmospheric moisture cannot be excluded. Consequently, the absolute BC δ¹⁸O values reported here may include a preparation-related component and should be interpreted cautiously.

5. Conclusions

Modern pig bone preserves complex isotopic signatures because skeletal tissues integrate dietary, hydrological, geological, and management-related inputs over time. In this study, ⁸⁷Sr/⁸⁶Sr ratios in bone closely overlapped with those in manufactured feed but differed significantly from those in groundwater. These findings suggest that Sr incorporation in pigs raised in intensive production systems is probably influenced more by dietary inputs rather than by local water or bedrock-derived signatures. However, bone ⁸⁷Sr/⁸⁶Sr values extending beyond the measured feed range indicate that the sampled feed alone cannot account for the full range of skeletal Sr isotope signatures. This suggests contributions from unmeasured or temporally variable sources, including differences among mineral supplements, groundwater, incidental soil or dust ingestion, and Sr incorporated earlier in life. Because the intake amounts and bioavailable Sr fluxes required for source apportionment were not determined for feed and groundwater, their relative contributions to skeletal Sr could not be quantified using a concentration-weighted mixing model. Thus, the observed overlap in ⁸⁷Sr/⁸⁶Sr values is insufficient to quantitatively establish feed as the dominant source of skeletal Sr. Consequently, skeletal Sr isotopes alone have limited utility as simple geographic tracers in modern livestock systems, particularly when feed ingredients are sourced beyond the rearing region. Future studies should quantify Sr concentrations, source-specific intake, and Sr bioavailability to enable concentration-weighted mixing analyses and more robust estimates of the relative contributions of these sources to skeletal Sr.

Notably, δ¹³C_BC and δ¹⁵N_BC values retained dietary information and exhibited enrichment compared to δ¹³C_Feed and δ¹⁵N_Feed values. These isotopic patterns reflect the incorporation of mixed plant-based diets and effects of metabolic fractionation. Oxygen isotopes exhibited the strongest hydrological association: Both δ¹⁸O_BC and δ¹⁸O_BP were significantly correlated with drinking-water δ¹⁸O, with δ¹⁸O_BP explaining a greater proportion of variance than δ¹⁸O_BC. The systematic offset between the two tissues suggests a coordinated, but tissue-specific, incorporation of oxygen in the skeletal system.

Collectively, these findings suggest that no single isotope system is sufficient to determine pig provenance under modern farming conditions. Instead, provenance assessment requires an integrated multi-isotope framework in which Sr isotopes reflect feed-mediated geological mixing, carbon and nitrogen isotopes constrain dietary inputs, and oxygen isotopes preserve information related to regional water sources. To the best of our knowledge, this study is the first to establish a skeletal multi-isotope baseline for modern pigs in South Korea. It provides an important reference framework for ecological and provenance mapping investigations of faunal systems.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c07049.

• Procedural blanks; isotope normalization and correction; reference materials used for isotope normalization (Table S1); representative chromatograms (Figures S1 and S2) (PDF)

#.

M.K.G. and S.-H.C. contributed equally to this work. M.K.G.: conceptualization, methodology, visualization, writingoriginal draft. S.H.C.: conceptualization, methodology, investigation, formal analysis, visualization, writingoriginal draft. W.J.S.: conceptualization, investigation, visualization, writingreview and editing. S.P.: investigation, formal analysis. H.-B.C.: investigation, formal analysis. K.-S.L: conceptualization, investigation, methodology, writingreview and editing.

The authors declare no competing financial interest.