3.1. Physico–chemical properties of soil and sediment samples
Several physicochemical parameters and statistics for the studied soils and sediments are provided in Table 2. The mean pH in the farm soils (FS) samples and river sediments (RS) were determined to be 6.85 and 7.47, respectively. Although both substrates were found to be in the neutral pH range, the RS typically had higher pH values (≥ 7.0) while the FSs were typically below 7.0. While the average silt fraction was typically higher in the FS (above 40%) than in the RS (below 20%), on average, the RS were comprised of more sand fraction (76%) than the FS (53%). In all samples, the clay content was below 40%, but still significantly higher in the FS than in RS (p < 0.05). The highest soil organic matter (SOM) content was observed at site FS1 with 10.4%. The highest SOM in RS was observed at site RS4 (1.98%) while the lowest was recorded at sites RS1 and RS5 (0.8%). FS markedly exceeded RS in terms of SOM. Electrical conductivity (EC) values for FS averaged around 518 µS cm− 1 while for RS, this value dropped to 238 µS cm− 1. The EC values are indicative of the levels of dissolved salts and nutrients in the substrates and hence, points to higher levels of nutrients in the soil over the sediments. This is further highlighted by the higher abundance of macronutrients (Mg, K and Ca) in the FS compared to the RS.
3.2. Heavy metal concentrations in farm soils and river sediments
Metal concentrations in 20 farm soil samples (FS1–FS20) from Matanagata village are shown in Table 6. Metal concentrations followed the order Fe (5.11%) > Mn (503.8 mg kg− 1) > Cr (136.2 mg kg− 1) > Cu (110.4 mg kg− 1) > Zn (80 mg kg− 1) > Ni (57.2 mg kg− 1) > Pb (45.4 mg kg− 1) > Co (35.4 mg kg− 1) > Cd (1.7 mg kg− 1). The concentrations of the studied metals showed little variations at all sites except for Pb, Cu, and Zn. The highest Zn concentration (116.85 mg kg− 1) and Mn concentration (544.9 mg kg− 1) were observed at site FS7 while the highest values for Ni and Cu (63.5 and 131.2 mg kg− 1) were both observed at site FS12. The highest concentration of Pb (63.64 mg kg− 1) was recorded at site FS20. An increasing trend in soil Pb concentration was observed moving downstream away from the mine with the farm sites further downstream (FS13–FS20) having the highest levels of Pb. However, the opposite case was noted for Zn and Fe, as soil concentrations generally decreased with distance from the mine site pointing to possible enrichment from the mine. The highest concentrations of Cd, Cr, Mn and Fe were also observed in farm sites with closest proximity to the mine tailings.
The concentration of heavy metals in Nasivi river sediments are summarized in Table 3 and designated RS1 to RS9. The range of metals in the sediments was: Pb (59.3–77.6 mg kg− 1), Zn (53.9–80.7 mg kg− 1), Cd (1.6–2.1 mg kg− 1), Ni (40.4–62.6 mg kg− 1), Cu (59.1–123.2 mg kg− 1), Cr (91.2–205.9 mg kg− 1), Mn (416.0–643.5 mg kg− 1), Co (24.5–42.7 mg kg− 1), Fe (3.84–5.58%). For all metals except Mn, the highest average concentrations were found in the midstream section of the Nasivi river, closest to the mine tailing dams. The highest concentrations of Zn, Cd, Ni, Cu, and Co were all observed at Site RS24, while the highest Fe and Cr concentrations were found at site RS25. Metal concentrations at sites sampled upstream before the mine generally had lower metal concentrations compared to sites midstream and downstream. This was confirmed by statistical analyses which revealed that heavy metal contamination differed between the upstream and both the midstream and the downstream sections of the river (P < 0.05). Site RS27, located downstream from the VGML, was found to contain the lowest concentration of Zn, Cd, Ni, Cr, Co and Fe amongst all sites. This decrease in metal concentration downstream may be due to ‘dilution’ by local sediment supply, chemical sorption, dissolution of contaminants and/or contaminant uptake by biota (Hudson-Edwards et al., 2008).
3.4. Comparison of soil and sediment metal concentrations at Vatukoula with international standards and background concentrations.
Table 4 presents the metal concentrations from FS in comparison to standard guidelines and permissible limits established by international bodies including the WHO Maximum Permissible Concentration (MPC), Canadian Agricultural Soil Quality Criteria (CASQC), Dutch Target Value (DTV) and background metal concentrations for global soils obtained from (Kabata–Pendias, 2010). Pb, Mn and Zn concentrations in all studied soils were found to be below the 100 mg kg− 1 permissible limit set by the WHO and also below the CASQC and DTV. However, although Cd values were below the WHO limit (3 mg kg− 1), it exceeded both the CASQC (1.4 mg kg− 1) and DTV (0.8 mg kg− 1) guidelines by 24.3% and 117.5%, respectively. For Co, the average concentration was above the WHO MPC (30 mg kg− 1) but below the CASQC (40 mg kg− 1). Mean concentrations of Pb, Cd, Ni, Cu, Cr, Co and Fe were found to be 68%, 14.3%, 324.4%, 97%, 183.8%, 130.8%, 3%, 18% and 89.3% above background levels, respectively. The average Cu (110 mg kg− 1) and Cr (136 mg kg− 1) concentrations in soil were both found to exceed all three guidelines as well as the natural background levels for global soils, elucidating possible enrichment from the mine (Kabata–Pendias, 2010). Although both Cu and Cr are considered essential micronutrients, they have been widely reported to display toxicity at levels that exceed cellular homeostatic control capacity (Sheehan et al., 1991; Stern, 2010). In fact some scientists regard Cu as the most toxic essential element (Hunter, 2015), while Cr, particularly in its hexavalent state (Cr[VI]), has been classified as a human carcinogen by the WHO.
On the other hand, Nasivi river sediments had higher average values for Pb, Cd and Cr compared to the FS, while the concentrations of Zn, Ni, Cu, Co, and Fe were generally lower. For reference, sediment concentrations were compared against Canadian Freshwater Sediment Quality Criteria (CFSQC) and the Australian and New Zealand Environment and Conservation Council (ANZECC) guidelines as shown in Table 4. Average sediment concentrations of all metals except Zn, Mn, Co and Fe, were found to greatly exceed both CFSQC and ANZECC guidelines. Particularly notable were the concentrations of Pb, Cd, Cu and Cr which exceeded CFSQC limits by 98%, 203%, 149.2% and 283.7%, respectively and ANZECC limits by 38.6%, 21.3%, 36.9%, 78.9%. For Ni, although no established CFSQC limit exists, the observed values exceeded the ANZECC guidelines by 136.5%. While the Nasivi river isn’t utilized as a drinking water source by the residents, livestock have been reported to drink from the river and these moderate levels may pose adverse risks. To properly contextualize the results, the baseline metal concentrations reported for the relatively ‘pristine’ Great Astrolabe reef in Fiji (Morrison et al., 1997), were used as local background values and compared with this study. All metals concentrations from the Nasivi RS greatly exceeded the concentrations reported for the Astrolabe reef. Concentrations of Pb, Zn, Cd, Ni, Cu, Cr, Mn, and Co in Nasivi sediments were 812%, 137%, 90%, 471%, 201%, 616%, 145% and 279% higher than respective concentrations reported in the Great Astrolabe reef sediments (Morrison et al., 1997). Particularly concerning were the high relative difference in Pb, Ni and Cr concentrations at both sites. This strongly points to possible contamination of the sediments with heavy metals emanating from the mine tailing dams.
3.5. Comparison of Vatukoula soil and sediment heavy metal concentrations with similar sites locally and globally.
The concentrations for heavy metals in FS near the VGML are compared with those in similar areas locally and globally as shown in Table 5. Metal concentrations in Matanagata village soils were generally lower than the levels reported for road side soils in Fiji’s capital, Suva (Maeaba et al., 2019). However, the mean concentration of Ni, Cr, Co and Fe observed in this study exceeded those in Suva roadside soils. Similarly, the levels of Pb, Cu and Cr were greater than values reported in agricultural soils from the Avliyana antimonite mineralization area in Torul, Turkey (Sungur et al., 2020) and concentrations in paddy fields adjacent to the Co Dinh chromite mine, the largest mine in Vietnam (Kien et al., 2010). For sediments, it was observed that the concentrations of Pb, Zn, Cd and Cu from the Nasivi river were lower than the respective concentrations reported for the Suva Harbor in Fiji (Maata and Singh, 2008). While metal concentrations from this study were far below those reported for the Molonglo river in Australia (Marasinghe Wadige et al., 2016), they exceeded those reported for the Ipojuca river in Brazil (Silva et al., 2019)and the Jequetepeque River Basin in Peru (Yacoub et al., 2012).
3.6. Geochemical pollution assessment
3.6.1. Contamination Degree (CD) and Pollution Load Index (PLI)
The results for contamination degree (CD) and pollution load index (PLI) are presented in Fig. 3. Based on the results from the CD, it was clear that the FS all had CD values above 16, delineating these sites as having high degree of contamination. Most concerning was the CD values for the RS which were found to be above 32, making these sites very highly contaminated. Noteworthy were sites RS4 and RS8 which possessed the highest CD values amongst all sites. However, based on PLI analyses, all FS were considered unpolluted in regards to metals. In contrast, the RS were considered moderately polluted at sites RS1 and RS7, strongly polluted at site RS4, and significantly polluted at all other sites.
3.6.2. Geoaccumulation Index (Igeo)
A close observation of Fig. 4A shows that there was little variation in the range of Igeo values obtained for the FS. Amongst all heavy metal contaminants in this study, Cd had the highest Igeo values in soil. All the FS were considered moderately polluted with Cd with an average Igeo value of 1.5. Similarly, the Igeo values for Cu fell in the 1 ≤ Igeo < 2 class indicating moderate Cu pollution. For Zn, Mn and Co, Igeo values for the FS were found to be in the unpolluted class (Igeo ≤ 0), suggesting that these metals were not in excess of background levels. Similarly, the FS were considered unpolluted to moderately polluted with Pb, Ni and Fe (0 ≤ Igeo < 1). In contrast, the RS showed higher levels of contamination compared to the FS (Fig. 4B). The average Igeo values for Pb in sediments was 2.6 and fell in the moderately to strongly contaminated class (Igeo ≤ 3). This suggests that the Pb concentrations in the river system is an area of concern and may manifest toxicity. Similarly, the Igeo values for Cr and Ni in the sediments fell in moderately to heavily polluted class (2 ≤ Igeo < 3). Igeo values for Zn, Cd, Mn and Fe in the sediments were all ≤ 1, indicating no pollution while values for Co and Cu suggested moderate pollution. It is interesting to note that while Pb, Zn, Ni and Co were classified as unpolluted in FS, the Igeo levels in the sediments were much higher. This is indicative of possible heavy metal enrichment of the river system through flooding and leaching from the mine tailings.
3.6.3. Enrichment factor
The enrichment factors for the studied metals were determined to comment on the extent of contamination due to anthropogenic activities. The EF for each heavy metal was calculated relative to the background values after normalization with the element Fe and shown in Fig. 5. In FS, EF values for all metals except Cd were < 2 suggesting minimal enrichment of these elements. This indicates that the concentration of these elements in the FS are primarily of geogenic origin with little to no input from mining. As for Cd, the EF values observed for FS was > 2 indicating moderate enrichment likely from the VGML. In the sediments, there was higher enrichment of heavy metals compared to the soils. EF values for Zn, Cd, Cu, Mn, Co were typically < 2 and were considered minimally enriched. Additionally, the sediments were considered moderately enriched in terms of Ni and Cr which both had EF values in the 2 ≤ EF < 5 class. As for Pb, EF values revealed that the RS were significantly enriched with this metal. According to (Liu et al., 2016), an EF between 0.05–1.50 point to a natural origin for the heavy metal concentration, while an EF > 1.50 indicates possible enrichment from non–crustal materials, particularly anthropogenic sources. On this basis, it is safe to assume that Cd, Cu, Cr, Ni and Pb levels were likely enriched from mining activities at the VGML.
3.6.4. Ecological risk index
The ERI and PERI were introduced to determine the semi–quantitative evaluation of regional pollution level, according to the toxicity of heavy metals and the response of the environment. The ERI values for heavy metals in the FS were; Pb (4.81–11.79), Zn (0.91–1.67), Cd (94.70–144.50), Cu (11.71–16.87), Ni (8.13–10.95) Mn (0.91–1.12), Co (21.36–27.44) and Cr (3.40–6.13), while in RS the range were; Pb (40.58–52.89), Zn (1.83–2.73), Cd (48.99–65.46), Cu (9.98–20.80), Ni (23.20–35.99) Mn (2.01–3.11), Co (14.17–24.73) and Cr (9.12–20.59), respectively. In the FS, all metals except Cd presented low ecological risk (ERI < 40). ERI values for Cd indicated considerable ecological risks. However, in the RS, only Pb and Cd were found to present moderate ecological risks (40 ≤ ER < 80) while other metals presented low risks (ERI < 40). Generally, Pb and Cd were the main contributors to the potential ecological risk in the RS. Several studies have suggested that mining activities generate Cd contamination which can seriously threaten the growth of marine life and human health due to its intense biotoxicity and high bioaccumulation potential (Liu et al., 2019; Zhou et al., 2018).
3.7. Human health risk assessment
As widely emphasized, the accrual of heavy metals in the tissues of biological organisms can lead to toxicity and induce several disorders. Heavy metals do not only compete with essential elements due to their chemical similarities but they also interact with several divalent transporters which affects various physiologic functions including cardiovascular, neural, hematopoietic, immunological, and gastrointestinal systems, as well as a possible role in kidney dysfunction, anaemia, liver toxicity, cancer, and Alzheimer’s disease (Tchounwou et al., 2012; Jaishankar et al., 2014; Rehman et al., 2017). Therefore, the carcinogenic and non–carcinogenic health risk associated with heavy metal exposure from the three pathways (ingestion, inhalation and dermal contact) on adults and children were quantified using the chronic daily intake (CDI) and the hazard index (HI).
3.7.1. Non–carcinogenic risks
The non–carcinogenic risks associated with heavy metal exposure for both adults and children are presented in Table 6. The HI values for non–carcinogenic risks were higher for children when compared to adults. The HI values for adults were in the range of 4.21E–04 to 1.37E–01 while for children, the values ranged from 3.73E–03 to 2.19E + 00. For both adults and children, the HI values for heavy metals were found to be in the order of Co > Mn > Cr > Fe > Co > Pb > Cd > Ni > Cu > Zn. Heavy metals with HI values exceeding one (HI > 1) strongly suggest that the metal may induce non–carcinogenic effects. In the present study, the calculated adult HI values for all metals were below one (HI < 1) suggesting minimal non–carcinogenic risks for adults from exposure to metals in the area. However, it was observed that HI values for children, particularly for Co (2.19E + 00), Mn (2.19E + 00), Cr (1.42E + 00) and Fe (1.17E + 00) were all above one (HI > 1). These high values observed for children in the area indicates that heavy metal pollution, particularly Co, Mn, Cr and Fe, may pose some adverse health risks to children living in the area. For reference, Naz et al. (2016) reported elevated carcinogenic risks due to high levels of Fe, Mn, and Cr in drinking water near the Sukinda chromite mine in India. Mn is mainly neurotoxic and increased levels have been linked with neurodevelopmental effects in children from across the world (Grandjean and Landrigan, 2014; Coetzee et al., 2016). An assessment of children living in a Mn–mining region of Ukraine also observed impaired growth and rickets–like skeletal deformities in 33% of children (Duka et al., 2011). Similarly, high doses of cobalt may affect the heart, lungs, blood and thyroid (Paustenbach et al., 2013; Leyssens et al., 2017; Banza Lubaba Nkulu et al., 2018).
3.7.2. Carcinogenic risks
Heavy metals such as Pb, Cd, Cr, and Ni are considered category 1 heavy metals according to the International Agency for Research on Cancer (IARC, 1980). The carcinogenic risks (incremental lifetime cancer risk) associated with exposure to these heavy metals from different exposure pathways are provided in Table 7. The obtained results showed that the ILCR values ranged from 6.40E–07 to 1.58E–05 in adults and 5.90E–06 to 1.46E–04 in children. According to US–EPA recommendations, cancer risks exceeding 1 in 10,000 exposure (ILCR > 1.00E − 04) are deemed unacceptable, while risks below 1 chance in 1,000,000 lifetime exposure (ILCR < 1.00E − 06) are not expected to pose significant health effects, and risks between 1.00E − 04 and 1.00E − 06 are generally considered acceptable depending on the circumstances of exposure (Fryer et al., 2006; USEPA, 2014). While ILCR values for Pb, Cr and Ni for both adults and children were found to be well within the USEPA threshold risk limit (ILCR > 1.00E − 04), we observed that Cd posed the highest potential carcinogenic risk especially in children (ILCR = 1.46E–04). The ILCR values for heavy metals in decreasing order was Cd > Ni > Cr > Pb. For both adult and children, the ingestion route was found to be the largest contributor to ILCR. This emphasizes the need for constant monitoring of Cd in the soils and sediments surrounding the mine. Several studies have reported the presence of elevated Cd levels in farm soils and sediments located near mines and the adverse health risks (Galunin et al., 2014; Mohammed and Abdu, 2014; Zhou et al., 2018). Generally, the above results indicate that children have a slightly higher likelihood of experiencing adverse health effects due to Cd exposure around the Vatukoula area.
3.8. Apportionment of heavy metal sources
A Pearson correlation matrix (PCM) was performed to determine the relationships between soil physico–chemical parameters and heavy metals as shown in Table 8. The silt and clay fractions of the soils showed moderate but negative correlations with Pb and weak positive correlations with the metals Ni, Co and Fe, while for Zn, the soil fractions were moderately positively correlated. Clay particles usually have affinity for metal cations due to the negative charges and are thought to adsorb metal ions through both ion exchange and specific adsorption (Rieuwerts et al., 1998; Bradl, 2004). Soil pH showed a moderate positive correlation with Pb; however, metals such as Ni, Cu, Co and Fe were all showed weak negative correlations with the pH. This is in contrast with most studies which established that the adsorption and mobility of cationic metals increased linearly with pH (Harter, 1983; McLean and Bledsoe, 1992; Barrow and Whelan, 1998); however, it has been reported that the retention of the metals did not significantly increase until the pH was greater than 7 (Harter, 1983). As for soil EC and WHC, we observed that both variables presented a moderately positive correlation with Ni, Cu and Co, while for Pb, the correlation was inverse. Similarly, SOM and CEC presented a moderately positive correlation with Zn, Ni, Cu, Co and Fe while inversely correlated with Pb and Cd. Amongst heavy metals, a particularly strong positive correlation was observed between Co, Fe and Ni. In addition, Fe was inversely correlated with Pb but positively correlated with Zn, and Cr. With regards to Mn, Pb, and Cd, these metals did not particularly show meaningful correlations with other metals in the study.
A principal component analysis (PCA) was used to deduce the sources and major contributors to metal pollution in the soils and sediments around the Vatukoula mine. Based on the PCA analysis, three significant principal components (PCs) (eigenvalues ˃ 1) were identified, accounting for approximately 77.52% of the total percent variability as shown in Table 9 and Fig. 7. This strongly indicates that several controlling factors or sources were responsible for the heavy metal concentrations in the soil and sediments around the VGML. PC1 accounted for 41.26% of the total variance and was explained by the high loading for Co (0.489) and Ni (0.492), which was consistent with the high correlation among these metals (r = 0.888). Moreover, Cu, Zn and Fe were also added into PC1 considering their close correlations with Co and Ni. Based on the low concentrations of these metals observed in this study, it is inferred that PC1 components are likely derived from natural sources. The metal relationships obtained from the PCA is corroborated by the correlation results (Table 8) in which Zn, Ni, Co and Fe presented moderate to strong positive correlations thus, indicating their common geogenic/natural sources. PC2 accounted for 20.05% of the total variance associated with Pb, Cd and Cr. Considering the fact that Pb has been associated with mining activities (Ogola et al., 2002; Rabiu et al., 2019), this suggests that metals associated with this component are likely sourced from the nearby mine. PC3, which comprised 15.86% of the contribution, had loadings of Cu (0.478), Mn (0.546), Fe (–0.461) and Cr (–0.355).