Overview
This study characterizes mercury (Hg) speciation in water and river sediment samples from two artisanal small-scale gold mine (ASGM) sites in the Kakamega and Vihiga gold belt, western Kenya (Rosterman and Malinya mine sites). It has two components: (1) development and validation of a novel dithizone-functionalized solid-phase extraction (SPE) cartridge method for field collection and preservation of Hg2+ and methylmercury (MeHg) in water samples, achieving representative recoveries over up to four weeks; and (2) application of this method alongside established HPLC-ICP-MS and DMA-80 techniques to characterize Hg speciation and co-occurring potentially toxic elements (arsenic, chromium, nickel, cadmium, lead) in ASGM-impacted water and sediment.
Water MeHg concentrations were below the detection limit (less than 0.007 ug/L) at both mine sites. Inorganic mercury (Hg2+) in waters ranged from 0.06 to 0.67 ug/L, below the WHO drinking water guideline of 1 ug/L total Hg; however, drinking water sources contributed 20 to 30% of the tolerable weekly intake of Hg (0.7 ug/kg body weight/week), indicating a need to minimize exposure from dietary sources. Sediment total Hg concentrations ranged from less than 0.05 to 1.84 mg/kg, with 11 of 21 samples exceeding the US EPA sediment guideline of 0.18 mg/kg total Hg. Sediment MeHg concentrations ranged from less than 0.1 to 36.8 ug/kg, with the highest values associated with stagnant ponds favorable to bacterial methylation. Chromium, arsenic, and nickel in sediments substantially exceeded US EPA freshwater sediment screening benchmarks at nearly all sampling points.
This paper is an environmental occurrence and methods paper, not a food occurrence paper. It provides context for Hg contamination pathways in ASGM-affected river systems, including drinking water sources and sediments that may enter local food chains via aquaculture, irrigation, and direct water consumption.
Study design
Cross-sectional field campaign at two ASGM sites in the Kakamega gold belt, western Kenya (Rosterman mine: 0 degrees 15’ 35.6” N, 34 degrees 43’ 12.0” E; Malinya mine: 0 degrees 11’ 22.4” N, 34 degrees 44’ 10.3” E). The rivers sampled are small tributaries of the River Yala and Nzoia, which ultimately drain into Lake Victoria. Twenty-five water samples (including preserved Hg SPE-cartridge samples) and twenty-one sediment samples were collected at intervals along the river passing through each mine site, ore washing ponds, disused alluvial mining areas, and drinking water sources. Sampling points (26 labeled Hg-001 to Hg-026) were distributed across both sites, with 21 at Rosterman and 5 at Malinya (Table 3, Table 5; Fig. 4 of source).
Samples
Total samples: 25 water (n = 25 for Hg speciation; additional multi-element analysis) and 21 sediment samples. No replicate field collections are described; sample sizes are as collected.
Water sample types: river water at various distances from mine activities, ore washing pond water, disused alluvial mining pond water, drinking water reservoir, underground drinking water pump water, mine water used as drinking water.
Sediment sample types: river bed sediment along mine river corridors, sediment from ponds formed by drinking water infrastructure, sediment from mine tailings areas.
Matrices: water (river, pond, drinking water sources) and river/pond sediment. Food matrices not measured.
Methods
Hg speciation in water (SPE-HPLC-ICP-MS): Dithizone-functionalized C18 SPE cartridges (Bond Elut Jr, Agilent) prepared by sequential washing with methanol, sodium formate-formic acid buffer (0.5 M, pH 4), 0.002 M dithizone solution, and buffer wash steps. Field water samples (30 mL, filtered 0.45 um HDPE) passed through cartridges at approximately 10 mL/min; cartridges stored at 4 degrees C in the dark until elution with 30 mL 2-mercaptoethanol solution (1% v/v). Eluent analyzed by HPLC-ICP-MS: aqueous-C18 column (Poroshell, 5 um, 4.6 mm x 50 mm, Agilent), mobile phase 0.5% v/v 2-mercaptoethanol, Agilent 1290II pump coupled to Agilent 8900 ICP-MS. Mass 201Hg used for Hg species quantification in SPE-water method. Method detection limits for Hg199 and Hg201 were determined as 3 times the standard deviation of 9 blank replicates (Supplementary Table 2 of source; exact numeric values not in main text). MeHg water detection limit stated in results as less than 0.007 ug/L MeHg.
Total Hg in sediment (DMA-80): 0.05 g sediment portions weighed into metal sample boats and subjected to thermal decomposition and direct Hg quantification on a DMA-80 total Hg analyzer (Milestone). Certified reference materials HR-1 (NRCC) and TH-2 (NRCC, Canada) analyzed every 10 samples to determine accuracy and precision (Supplementary Table 4).
MeHg in sediment (isotope dilution HPLC-ICP-MS): Isotopic spike of Me(199Hg) added to 0.5 g sediment samples, equilibrated, then extracted with 10 mL of 3 M HCl / 0.2 M citric acid solution by microwave (500 W, 5 min, 45 degrees C). Solution filtered and analyzed by HPLC-ICP-MS using masses 199Hg and 201Hg. Concentration calculated by isotope dilution formula. Estuarine certified reference material EU ERM-CC58 analyzed alongside samples (Supplementary Table 4).
Multi-element analysis in water and sediment (ICP-MS): Water samples preserved by acidification with 0.5% HCl and 1% HNO3 and analyzed by ICP-MS (Agilent 8900, British Geological Survey). Sediment samples subjected to mixed acid digestion (HF/HNO3/HClO4) using programmable hot-block at British Geological Survey (Watts et al. 2008, 2013, 2019 protocol). CRMs USGS BCR-2, NRCC LKSD-1, NRCC LKSD-3, and NRCC MESS-4 analyzed alongside sediment samples; CRM recoveries greater than 85% and standard deviation less than 10% (Supplementary Table 6). Duplicate sample analysis demonstrated percentage differences less than 25% for all assessed elements.
SPE method validation: Synthetic river water spiked with 0.5 ug/L Hg2+ and 0.05 ug/L MeHg (n = 5 per condition) tested across pH 5 and 8, chloride concentrations 0.1% and 1.9% w/v, and humic acid concentrations 5 and 25 mg/L. All conditions achieved greater than 84% recovery of both species (Table 2). Four-week stability: cartridges loaded with spiked solution stored 0, 1, 2, 3, and 4 weeks in refrigerated (4 degrees C, dark) and unrefrigerated (16 degrees C, dark) conditions.
Instrument
HPLC-ICP-MS: Agilent 1290II pump coupled to Agilent 8900 ICP-MS. DMA-80 total Hg analyzer (Milestone). ICP-MS (Agilent 8900, British Geological Survey) for multi-element analysis.
LOD/LOQ
Water MeHg detection limit: less than 0.007 ug/L MeHg (stated in results text as the threshold below which all water samples fell). Method detection limits for Hg199 and Hg201 by HPLC-ICP-MS were determined as 3x standard deviation of 9 blank replicates; numeric values reported only in Supplementary Table 2 of the source and are not reproduced in the main text. Sediment MeHg detection limit: less than 0.1 ug/kg MeHg (stated in results for samples below detection). No LOQ values reported in main text. No LOD/LOQ reported for multi-element ICP-MS analysis in main text.
CRM recovery
Total Hg in sediment (DMA-80): CRMs HR-1 and TH-2 (NRCC Canada) analyzed every 10 samples; numeric recoveries in Supplementary Table 4 only. MeHg in sediment (isotope dilution): estuarine CRM EU ERM-CC58 extracted and analyzed alongside samples; numeric recoveries in Supplementary Table 4. Multi-element sediment (ICP-MS): CRMs USGS BCR-2, NRCC LKSD-1, NRCC LKSD-3, NRCC MESS-4; stated recoveries greater than 85% with standard deviation less than 10% (Supplementary Table 6). Aqueous Hg QC: a 0.25 ug/L Hg quality control solution and a 0.04 ug/L MeHg quality control solution analyzed every 10 samples during HPLC-ICP-MS analysis; numeric results in supplementary materials only.
Results values
Water Hg speciation (Table 3):
MeHg in water: below detection limit (less than 0.007 ug/L MeHg) at all 25 sampling points across both Rosterman and Malinya mine sites.
Hg2+ in water, Rosterman mine (21 sampling points, Hg-001 to Hg-021): river water leaving mine site (Hg-001 to Hg-003) 0.059 to 0.106 ug/L; river water used for domestic activities (Hg-004) 0.229 ug/L; river water after mixing point (Hg-005) 0.272 ug/L; mixing point with minor tributary (Hg-006) 0.405 ug/L; minor tributary from nearby villages (Hg-007) 0.165 ug/L; Rosterman mine river prior to mixing (Hg-008) 0.058 ug/L; river upstream of drinking water reservoir (Hg-009) 0.119 ug/L; river downstream of mining activities (Hg-010) 0.090 ug/L; river downstream of drinking water reservoir (Hg-011) 0.105 ug/L; small pond from drinking water reservoir outlet (Hg-012) 0.059 ug/L; drinking water reservoir (Hg-013) 0.055 ug/L; river downstream of disused ore washing pond (Hg-014) 0.103 ug/L; river immediately after disused alluvial mining pond (Hg-015) 0.062 ug/L; disused alluvial mining pond (Hg-016) 0.095 ug/L; river prior to disused alluvial mining pond / drinking water collection point (Hg-017) 0.315 ug/L; river prior to entering mine site (Hg-018) 0.293 ug/L; ore washing pond (Hg-019) 0.075 ug/L; mine water used as drinking water (Hg-020) 0.144 ug/L; ore washing pond (Hg-021) 0.671 ug/L (highest at Rosterman site).
Hg2+ in water, Malinya mine (5 sampling points, Hg-022 to Hg-026): river water exiting mine site near alluvial working point (Hg-022) 0.216 ug/L; river water exiting mine site (Hg-023) 0.115 ug/L; drinking water from underground pump (Hg-024) 0.413 ug/L; pond from drinking water pump (Hg-025) 0.606 ug/L; river flowing through mine tailings (Hg-026) 0.574 ug/L.
WHO drinking water guideline for total Hg: 1 ug/L (WHO 2005). All water samples below this guideline. US EPA chronic water quality criterion: 0.77 ug/L total Hg (EPA 1995). All water samples below this limit.
Drinking water exposure assessment (Table 4): Tolerable weekly intake 0.7 ug Hg/kg body weight/week (WHO 2003 as cited in source). Assuming 60 kg body weight and 3 L/day water consumption (worst-case worker exposure), Hg2+ consumption from the underground drinking water pump at Malinya (Hg-024) was 0.15 ug/kg body weight/week (21% of TWI). The pond from the drinking water pump (Hg-025) contributed 0.21 ug/kg body weight/week (30% of TWI). Rosterman drinking water sources contributed 4 to 17% of TWI. At 2 L/day average consumption, contributions were 4 to 11% of TWI at Rosterman and 20 to 31% of TWI at Malinya drinking water sources.
Co-occurring elements in water (Table 3): total chromium (Cr) 0.09 to 2.07 ug/L (below WHO guideline 50 ug/L); total arsenic (As) 0.14 to 11.2 ug/L (mostly below WHO 10 ug/L guideline; highest at ore washing ponds Hg-019 at 10.4 to 11.2 ug/L); total nickel (Ni) 1.09 to 17.6 ug/L (below WHO 70 ug/L); total cadmium (Cd) mostly below 0.006 ug/L at Rosterman; 0.007 to 0.032 ug/L at Malinya; total lead (Pb) mostly below 0.02 ug/L at Rosterman; 0.03 to 0.06 ug/L at Malinya.
Sediment Hg speciation (Table 5):
Total Hg in sediment: Rosterman mine, less than 0.05 to 1.72 mg/kg (highest at disused alluvial mining pond Hg-015 at 1.72 mg/kg and Hg-017 at 1.01 mg/kg); Malinya mine, less than 0.05 to 1.84 mg/kg (highest at pond from underground drinking water pump Hg-025 at 1.84 mg/kg). Overall range 0.08 to 1.84 mg/kg (abstract states 0.08 to 1.84 mg/kg). 11 of 21 sediment samples exceeded the US EPA guideline of 0.18 mg/kg total Hg.
MeHg in sediment: Rosterman mine, less than 0.1 to 8.2 ug/kg (highest downstream of drinking water reservoir, Hg-011 at 8.2 ug/kg; Hg-012 pond 3.3 ug/kg; Hg-014 downstream of disused ore washing pond 4.9 ug/kg; Hg-015 immediately after disused alluvial mining pond 0.9 ug/kg; Hg-016 disused alluvial mining pond 3.8 ug/kg; Hg-009 upstream of drinking water reservoir 2.8 ug/kg). Malinya mine, 1.0 to 36.8 ug/kg (highest at pond from underground drinking water pump Hg-025 at 36.8 ug/kg; Hg-022 2.2 ug/kg; Hg-023 6.1 ug/kg; Hg-026 through mine tailings 1.0 ug/kg). Mean sediment MeHg across all sites: 5.9 plus or minus 14.3 ug/kg (stated in abstract).
Co-occurring elements in sediment (Table 5): total Cr 275 to 1940 mg/kg (Rosterman 482 to 1940 mg/kg; Malinya 275 to 316 mg/kg); US EPA guideline 43.4 mg/kg — all 21 sediment samples exceeded this limit. Total Ni 64.2 to 358 mg/kg; US EPA guideline 22.7 mg/kg — all 21 samples exceeded. Total As 7.94 to 349 mg/kg (Rosterman 7.94 to 77.4 mg/kg; Malinya 28.5 to 345 mg/kg); US EPA guideline 9.8 mg/kg — 19 of 21 samples exceeded. Total Cd 0.04 to 0.32 mg/kg; US EPA guideline 0.99 mg/kg — no samples exceeded. Total Pb 15.9 to 43.4 mg/kg; US EPA guideline 35.8 mg/kg — 3 samples exceeded (Hg-010 at 41.5 mg/kg, Hg-001 at 15.9 mg/kg noted in text, and Hg-025 at 38.2 mg/kg).
SPE validation (Table 2): Recovery of Hg2+ and MeHg from spiked synthetic river water (n = 5 per condition) across matrix conditions: pH 5 (Hg2+ 100 plus or minus 1%, MeHg 90 plus or minus 1%); pH 8 (Hg2+ 98 plus or minus 7%, MeHg 84 plus or minus 7%); 0.1% Cl (Hg2+ 101 plus or minus 3%, MeHg 98 plus or minus 4%); 1.9% Cl (Hg2+ 99 plus or minus 1%, MeHg 96 plus or minus 4%); 5 mg/L humic acid (Hg2+ 100 plus or minus 1%, MeHg 92 plus or minus 6%); 25 mg/L humic acid (Hg2+ 105 plus or minus 3%, MeHg 91 plus or minus 5%). All conditions greater than 84% recovery with standard deviation less than 8%.
Four-week preservation: mean recovery of MeHg over 4 weeks was 109 plus or minus 13% unrefrigerated and 115 plus or minus 8% refrigerated; Hg2+ recovery was 93 plus or minus 12% unrefrigerated and 100 plus or minus 14% refrigerated. At 4 weeks specifically: 96 plus or minus 14% Hg2+ and 110 plus or minus 12% MeHg.
Evidence fitness
This source provides direct environmental occurrence measurements for total Hg and MeHg in water and sediment from ASGM-affected river systems in Kenya, with multi-element co-contamination data (Cr, As, Ni, Cd, Pb) for the same matrices. It does not measure food matrices. Its value for this wiki is as supply-chain and exposure-pathway context for Hg contamination near ASGM sites that may affect local food systems (aquaculture in Lake Victoria tributaries, irrigation water, aquatic food sources). The SPE method development component is methodological, not an occurrence study. Evidence fitness for food occurrence claims: context only. Evidence fitness for environmental pathway characterization near ASGM sites: strong, with well-documented methods and CRM quality control.
Limitations
Water sampling was conducted at a single time point; Hg concentrations in ASGM-affected rivers may fluctuate substantially with alluvial mining activity intensity and seasonal conditions, as the authors note. The sampling year is not stated in the main text. No fish or aquatic food organism tissue measurements are included despite the aquaculture context noted in the conclusion. Exact numeric method detection limits for the HPLC-ICP-MS water analysis are reported only in Supplementary Table 2 and are not reproduced in the main text. CRM recovery statistics for total Hg and MeHg in sediment are in Supplementary Table 4 only. The tolerable weekly intake calculation assumes a 60 kg average Kenyan body weight and fixed water consumption rates; actual exposure variability across individuals is not characterized. Hg speciation in sediments reports total Hg by DMA-80 and MeHg by isotope dilution but does not report Hg2+ speciation separately in sediment. Arsenic speciation (inorganic vs. organic species) is not performed; total As only is measured.
Regulatory context
WHO drinking water guideline for total Hg: 1 ug/L (WHO 2005). All water samples at both mine sites were below this guideline. US EPA chronic water quality criterion for total Hg: 0.77 ug/L (EPA 1995). US EPA freshwater sediment screening benchmarks (EPA 2006): total Hg 0.18 mg/kg; total Cr 43.4 mg/kg; total Ni 22.7 mg/kg; total As 9.8 mg/kg; total Cd 0.99 mg/kg; total Pb 35.8 mg/kg. Tolerable weekly intake for inorganic Hg: 0.7 ug/kg body weight/week (WHO 2003 as cited in source). No Kenyan national drinking water or sediment standards are cited in this paper.
Implications
Certification: This paper is not an occurrence study for food matrices. It provides supply-chain context for Hg and co-metal contamination in ASGM-impacted water and river sediment in Kenya. It is relevant to understanding the exposure pathways for communities near ASGM sites that consume locally sourced fish, irrigation-grown crops, or contaminated drinking water. No direct certification threshold implications.
Courses: Useful case study illustrating: (1) the distinction between total Hg and MeHg speciation in environmental monitoring; (2) how stagnant water and pond sediments promote bacterial methylation and produce elevated MeHg despite relatively low total Hg; (3) the contribution of ASGM activities to multi-metal contamination (Cr, As, Ni) substantially exceeding environmental screening benchmarks; and (4) how drinking water sources near ASGM sites can contribute a meaningful fraction (20 to 30%) of the weekly Hg tolerable intake even when total Hg is below the drinking water guideline.
Microbiome: Elevated iron concentrations (1250 to 1900 ug/L Fe) at the Malinya underground drinking water pump are linked to enhanced bacterial Hg methylation; the paper notes iron-reducing bacteria as key methylators. This iron and sulfate-reducing bacterial pathway is relevant to microbiome discussions of anaerobic Hg cycling.
App: No food matrix contamination data. Not applicable to ingredient contamination_profile blocks.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.