Co-occurrence of uranium, fluoride and nitrate in Sahelian groundwater of Agadez (Niger) drives population health risks

Abstract

This study investigates the co-occurrence of uranium (U), fluoride (F), and nitrate (NO₃⁻) in untreated groundwater used for drinking water in the city of Agadez, Niger, with implications for the wider Sahel region. Comprehensive hydrochemical analysis of all public and a cross-sectional sample of private drinking water wells revealed widespread contamination. Among drinking water wells, 63.6% exceeded the WHO guideline for uranium, 68.2% exceeded the fluoride guideline, and 13.6% exceeded the nitrate guideline. Deterministic health risk assessment using USEPA reference doses showed that 77.3% of drinking water wells pose cumulative non-carcinogenic health risks to adults and 86.4% to children (Hazard Index > 1, summed across U, F, and NO₃⁻). Risk levels were slightly higher for private than for public wells, indicating only a minor disparity in exposure across the population. Solute correlation analysis and saturation index calculations, supported by irrigation-well data, indicate distinct geochemical controls on contaminant occurrence. Uranium mobility in Paleozoic–Mesozoic sandstones of the Teloua Formation is enhanced by carbonate complexation in bicarbonate-rich waters, while fluoride released from mica minerals accumulates under alkaline conditions due to calcite–fluorite coupling and calcium depletion. In contrast, nitrate contamination is anthropogenic, with one private drinking water well exhibiting extreme concentrations (3,433 mg/L) linked to inadequate sanitation. These results indicate that a significant portion of Agadez’s 124,000 residents are chronically exposed to elevated levels of uranium, fluoride, or both through untreated groundwater sources. Given the regional continuity of similar geological formations and sanitation challenges, comparable multi-contaminant groundwater quality problems could occur across the Sahel. This study represents the first comprehensive assessment of uranium–fluoride–nitrate co-occurrence in Sahelian groundwater.

Introduction

In the Sahelian zone, where surface water bodies are ephemeral and precipitation is highly variable, aquifers constitute the primary source of water for domestic consumption, agricultural irrigation, and livestock watering^1,2. However, groundwater-dependent communities are at risk not only by climate variability and overexploitation but also by water contamination through both geogenic and anthropogenic sources^3,4.

The co-occurrence of geogenic and anthropogenic contaminants in groundwater presents significant challenges for water resource management in developing regions. Uranium (U), fluoride (F), and nitrate (NO₃⁻) represent three contaminants of particular concern in Sahelian aquifers, each posing distinct health risks^5,4.

The three contaminants pose distinct health risks. Uranium, a naturally occurring radionuclide, is nephrotoxic and causes chronic kidney disease through long-term ingestion even at low concentrations, prompting the WHO provisional guideline of 30 µg/L based on chemical rather than radiological toxicity^6–8. Fluoride causes skeletal and dental fluorosis above 1.5 mg/L, and the Sahel has been identified as a hotspot for groundwater fluoride contamination^9–11. Nitrate, mostly of anthropogenic origin, causes infant methemoglobinemia and is linked to chronic conditions in adults, including thyroid dysfunction and carcinogenic effects via nitrosamine formation^12,13. Beyond drinking water, both uranium and fluoride accumulate in irrigated crops, increasing dietary exposure^14,15.

Niger, located in the heart of the Sahel with the majority of its territory within the Sahara Desert, exemplifies the water quality challenges facing the region. Agadez is a historic city in the semi-arid northern part of Niger with approximately 124,000 inhabitants. Today, groundwater extracted from both the shallow alluvial aquifer and the deeper Teloua sandstone formation constitutes the sole source of drinking water for Agadez, serving private and public supply networks without any treatment prior to consumption.

Despite the critical importance of this groundwater resource, very few studies have comprehensively investigated water quality, including major ions, trace elements, and health-relevant contaminants. Previous water quality control by the Agadez Campus Centre identified high fluoride concentrations as the main water quality problem, with no further concerns mentioned regarding uranium or other contaminants.

Public health concerns regarding water quality emerged following reports from the local population of Agadez describing noticeable changes in dental coloration and an increased incidence of infections among infants. Subsequent chemical analyses of tap water revealed elevated concentrations of fluoride (F) in the municipal water supply. These findings prompted the local water authorities to initiate a comprehensive assessment of the hydrochemical characteristics of drinking water wells in order to evaluate the extent of fluoride contamination and its potential health implications which subsequently revealed co-occurring uranium and nitrate contamination.

Within the framework of technical cooperation between the Federal Institute for Geosciences and Natural Resources (BGR) and the Niger Basin Authority (NBA), a comprehensive groundwater quality investigation was conducted across the Agadez urban area in the years 2023 to 2024. Drawing from the resulting hydrochemical dataset – including samples from all public drinking water wells, a cross-sectional sample of private drinking water wells and some irrigation wells – the objectives of this study were to: (1) assess the concentrations of uranium, fluoride, and nitrate in groundwater used untreated as drinking water; (2) undertake a cumulative health risk assessment considering co-occurrence of multiple contaminants (U, F and NO₃⁻) in consumed groundwater for both adults and children; and (3) examine the hydrogeochemical processes responsible for contaminant mobilization and co-occurrence, with implications for geologically similar areas across the Sahel.

Study area

Geographic setting

The study area encompasses the urban center of Agadez and its surrounding agricultural zones in the northern, semi-arid Saharan region of Niger. Agadez is situated at the southern margin of the Air Massif (Fig. 1). The investigated groundwater wells are distributed within and around the urban area, with most drinking water wells located between the city center and the Teloua river. The Teloua river, a major ephemeral watercourse, originates in the mountainous terrain of the Air Massif at elevations around 850 m above sea level and drains southwestward through Agadez to an outlet elevation of approximately 430 m above sea level.

Fig. 1.

(a) Location map of the study area in Niger. (b) Satellite image of the study zone with the watershed boundary outlined in yellow and the main drainage network in blue; the red shaded box indicates the detailed mapping area. (c) Geological map of the study area around Agadez, displaying lithological units, major fault systems, and drinking water well locations.

Based on meteorological data from the German Meteorological Service (DWD), mean annual precipitation for the period 1961–1990 was 111 mm, increasing slightly to 151 mm for 1991–2009, reflecting a modest increase as a recovery from severe Sahelian droughts in the 1970s and 1980s¹⁶. Precipitation is highly seasonal, concentrated between June and September, with July and August receiving the majority of annual rainfall. Mean annual temperature in Agadez is approximately 31 °C, with monthly means ranging from around 22 °C in January to 37 °C in May, the hottest month preceding the onset of the rainy season.

Geological and hydrogeological framework

The city of Agadez is situated within the extensive Iullemeden sedimentary basin, which covers approximately 500,000 km² and was formed through depositional processes spanning from the Paleozoic to Tertiary periods¹⁷. The regional geology comprises several distinct lithological units that influence groundwater occurrence and chemistry (Fig. 1).

The crystalline basement, consisting of Precambrian granites, gneisses, and related metamorphic rocks, underlies the sedimentary sequence at depths of approximately 30 m beneath the city^18,19. Although the basement rocks are highly fractured, they are not extensively weathered in this area and are typically overlain by younger sedimentary formations.

The Teloua Formation represents the primary aquifer system exploited for water supply in the study area. This Upper Triassic to Lower Jurassic sandstone unit is lithologically heterogeneous, comprising fine to coarse sandstones interbedded with clayey sandstones, siltstones, and variegated (multi-coloured) clays. The total thickness of the Teloua Formation ranges from 28 to 80 m depending on location. Mineralogically, the Teloua sandstones are composed predominantly of quartz with subordinate K-feldspar (microcline) and Na-feldspar (albite), minor biotite and muscovite, hematite, and accessory apatite, monazite, and zircon²⁰. The presence of feldspar minerals is significant for understanding geochemical processes affecting water quality, as feldspar weathering releases sodium and can influence fluoride mobility¹⁰.

Quaternary alluvial and eluvial deposits fill topographic depressions and line the valleys of ephemeral streams. These unconsolidated sediments, composed of sands, gravels, and clays of variable thickness and lateral extent, host shallow groundwater that is readily accessible through shallow wells. The alluvial aquifers receive direct recharge from episodic flood events during the rainy season – a predominant recharge mechanism in the region^21,22.

Granite outcrops occur within the catchment area of the Teloua valley, representing exposures of the Air Massif plutonic complex²³. These granites, composed of biotite-bearing varieties and leucocratic phases, contribute to the regional hydrochemical evolution through weathering processes that release ions including fluoride from mica minerals. Uranium-bearing minerals in the region are primarily associated with the Paleozoic–Mesozoic sandstones of the Teloua Formation and related units of the Tim Mersoï Basin, as well as with granitic rocks of the Aïr Massif, which are known sources of geogenic uranium through weathering and groundwater–rock interaction^24,25.

Wells sampled in this study are predominantly completed in two hydrogeological settings: the shallow alluvial/eluvial aquifers within the Teloua valley catchment, and the deeper Teloua sandstone formation.

Materials and methods

Groundwater sampling and analysis

A total of 30 groundwater wells were sampled across the Agadez urban area and its surrounding agricultural zones during the dry season (April–May 2024). The wells exhibit an average depth below ground level of circa 70 m. The 30 wells included 22 wells used for drinking water and 8 wells used for irrigation. Of the drinking water wells, 10 were public wells managed by local authorities and 12 were privately operated household wells. The 10 public drinking water wells represent a complete coverage of all public wells supplying drinking water to the population within the Agadez urban area. Only one public well located approximately 75 km outside the city limits was not sampled. The irrigation wells were all privately operated.

Private wells in Agadez are numerous and no complete or reliable inventory exists. Consequently, the sampled private wells represent a non-exhaustive but cross-sectional sample, selected to capture the range of groundwater quality conditions encountered in privately operated drinking and irrigation wells across the city.

All wells were in service at the time of sampling, and water samples were collected only after field parameters (pH, electrical conductivity, and temperature) had stabilized. Field measurements were recorded using a calibrated WTW 350i multimeter.

Samples were collected in 100 mL and 500 mL bottles. For major cation analysis, samples were filtered directly in the field using 0.45 μm cellulose acetate filters connected to a syringe and subsequently acidified with 69% HNO₃. Separate unfiltered samples were collected for the determination of anions and alkalinity.

Chemical analyses were conducted at the BGR water analysis laboratory in Hannover, Germany. Major cations (K⁺, Na⁺, Mg²⁺, Ca²⁺, Fe²⁺, Mn²⁺) were analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES, SPECTRO ARCOS) with limits of quantification (LOQ) below 0.04 mg/L. Anions (Cl⁻, SO₄²⁻, NO₃⁻, F⁻, Br⁻) were determined by ion chromatography (Thermo Scientific Dionex ICS-5000+) using a guard and analytical column (Dionex IonPac AG19/AS19) combined with a membrane suppressor technique. The LOQ for all anions was below 0.03 mg/L. Uranium and other trace elements including Arsenic were determined by inductively coupled plasma–mass spectrometry (ICP-MS) with an ESI prepFAST autosampler in the inorganic geochemistry laboratory of the BGR. The individual element concentrations were determined using standard calibration curves (linear regression). For quality and plausibility control of the measurements, internal element standards were added to each sample, and between approximately every 10–15 samples an external reference standard (TMDA 62.2 – trace metal fortified river water) was measured for quality control purposes.

Alkalinity was determined by potentiometric titration of 50 mL unfiltered sample aliquots with 0.05 N HCl to a pH 4.3 endpoint using a Mettler-Toledo T90 autotitrator. Analytical uncertainties were ± 5% for all anions and cations. Charge balance errors calculated for all water samples were within ± 3%, confirming the analytical quality. Since all 30 samples met this criterion, no samples required exclusion; any sample exceeding the ± 5% threshold conventionally applied in hydrochemical studies would have been excluded from interpretation²⁶.

Saturation indices (SI) for calcite, fluorite, and schoepite were calculated from measured water chemistry using ion activities derived from ionic strength, applying SI = log(IAP/K) with equilibrium constants at 25 °C and accounting for pH-dependent uranyl–carbonate complexation in the case of schoepite.

Statistical analysis and outlier detection

Statistical analysis was performed to evaluate relationships between water quality parameters relevant to uranium, fluoride, and nitrate occurrence. Prior to correlation analysis, outliers were identified using the interquartile range (IQR) method, where values exceeding Q₃ + 1.5 × IQR or falling below Q₁ − 1.5 × IQR were flagged as potential outliers. One extreme outlier (sample AZ4 with NO₃⁻ = 3,433 mg/L) was excluded from correlation calculations to prevent distortion of statistical relationships, though this sample is discussed separately as it represents significant point-source contamination.

Pearson correlation coefficients (R) were calculated for the complete suite of major ions, trace elements, and field parameters using samples with complete analytical data excluding one outlier (n = 29). Cross-correlations between U, F, and NO₃⁻ were examined specifically to test whether these contaminants share common geochemical controls or derive from independent processes. Statistical significance was assessed at the p < 0.05 level.

Health risk assessment

Human health risks associated with ingestion of untreated groundwater were assessed using the USEPA non-carcinogenic risk assessment framework²⁷, which is widely applied in similar studies^28–30.

Average daily dose (ADD, mg/kg/day) via drinking-water ingestion was estimated using the standard USEPA exposure equation, ADD = (C × IR × EF × ED)/(BW × AT), which reduces to ADD = (C × IR)/BW under the assumptions of continuous exposure (EF = 365 d/yr) and AT = ED × 365 d. Where C is the contaminant concentration (mg/L), IR is the water ingestion rate (2 L/day for adults; 1 L/day for children), and BW is body weight (70 kg for adults; 15 kg for children).

Non-carcinogenic risk was expressed as a hazard quotient, HQ = ADD/RfD. Reference doses were taken as: U = 0.003 mg/kg/day (IRIS oral RfD for soluble uranium salts), F = 0.06 mg/kg/day (USEPA IRIS oral RfD), and NO₃⁻ = 7.04 mg/kg/day (equivalent to 1.6 mg nitrate-N/kg/day).

For multi-contaminant screening, a hazard index (HI) was calculated as the sum of individual HQs. Because U, F, and nitrate affect different target organs, the summed HI is presented as a conservative screening indicator of cumulative chemical burden rather than an endpoint-specific hazard metric, an approach widely applied in multi-contaminant exposure studies^31–33,27. An HI > 1 indicates potential health risks from combined long-term exposure.

Results

Hydrochemical facies classification

The 30 groundwater samples exhibited temperatures ranging from 31 to 38 °C, and their electrical conductivity varied from 215 to 1832 µS/cm for most samples, with one exceptional value of 8950 µS/cm at well AZ4 indicating brackish conditions associated with severe anthropogenic contamination. The pH ranged from slightly acidic (6.2) to moderately alkaline (8.5), with 73% of samples having pH values above 7.0.

The Piper diagram (Fig. 2) reveals a clear hydrochemical evolution trend associated with increasing mineralization. Samples with the lowest EC values (< 400 µS/cm, shown in dark blue) cluster in the Ca-Mg dominated portion of the cation triangle and the HCO₃ dominated portion of the anion triangle, representing Ca-Mg-HCO₃ type waters. With increasing EC, samples progressively shift toward the Na + K apex in the cation triangle, while anion composition becomes more mixed with increasing contributions from SO₄²⁻ and Cl⁻. The most mineralized samples (EC > 1000 µS/cm, shown in orange and red) plot in the Na + K dominated field for cations and show variable anion compositions ranging from SO₄-dominated to mixed Cl-SO₄ types. This hydrochemical evolution from Ca-Mg-HCO₃ toward Na-SO₄-Cl facies is consistent with progressive water-rock interaction, cation exchange, and longer groundwater residence times. The implications of this calcium depletion for fluoride mobility are discussed in Sect.  5.2.

Fig. 2.

Piper diagram showing hydrochemical facies evolution of groundwater samples.

Uranium concentrations

A summary of contaminant concentrations and WHO guideline exceedances for all wells is provided in Table 1. Uranium concentrations in the 30 groundwater samples with uranium data ranged over four orders of magnitude, from below detection (0.03 µg/L at AZ13) to 533.7 µg/L at AZ9, with an overall median of 34.5 µg/L and mean of 76.8 µg/L. Drinking water wells showed a median of 49.0 µg/L, while irrigation wells exhibited a lower median of 20.3 µg/L (Fig. 3A). The distribution was strongly right-skewed, with several wells exhibiting extremely elevated concentrations.

Table 1.

Summary of contaminant concentrations and WHO guideline exceedances in drinking water wells. Nitrate range includes the extreme outlier AZ4 (3,433 mg/L); excluding AZ4, the drinking-well range is < 0.01–99.8 mg/L.

Contaminant	WHO guideline	Drinking wells (n = 22)	Irrigation wells (n = 8)
Uranium (µg/L)	30	Range 0.03–209.1	Range 4.4–533.7
		Median 49.0	Median 20.3
		14/22 (63.6%) exceed	3/8 (37.5%) exceed
Fluoride (mg/L)	1.5	Range 0.25–10.9	Range 0.39–4.2
		Median 4.40	Median 1.08
		15/22 (68.2%) exceed	2/8 (25.0%) exceed
Nitrate (mg/L)	50	Range < 0.01–3433	Range < 0.01–184
		Median 6.96	Median 5.79
		3/22 (13.6%) exceed	1/8 (12.5%) exceed

Fig. 3.

Comparison of uranium (A), fluoride (B), and nitrate (C) concentrations in drinking water wells versus irrigation wells.

Assessment against established thresholds revealed significant contamination particularly of water used for drinking purposes. Among drinking water wells (n = 22 with uranium data), 14 samples (63.6%) exceeded the WHO provisional guideline value of 30 µg/L, with the highest concentrations observed at AZ1 (209.1 µg/L), AZ18 (197.5 µg/L), and AZ6 (190.6 µg/L) – exceeding the guideline by factors of 6.4 to 7.0 (Fig. 3A). Among irrigation wells (n = 8), three samples (37.5%) exceeded the WHO drinking water guideline, with the highest concentration in the entire dataset observed at AZ9 (533.7 µg/L), an irrigation well located in the Teloua valley. This extreme value, nearly 18 times the WHO guideline, indicates localized elevated uranium mobilization.

Fluoride concentrations

Fluoride concentrations in the 30 groundwater samples with fluoride data ranged from 0.25 mg/L (AZ20) to 10.9 mg/L (AZ13), with an overall median of 2.59 mg/L and mean of 3.13 mg/L. Drinking water wells showed a median of 4.40 mg/L, while irrigation wells exhibited a lower median of 1.08 mg/L (Fig. 3B). Unlike uranium, fluoride concentrations did not exhibit extreme outliers, though the distribution remained positively skewed.

Comparison with established thresholds demonstrated widespread contamination across both water use categories. Among drinking water wells (n = 22), 15 samples (68.2%) exceeded the WHO guideline value of 1.5 mg/L (Fig. 3B). Wells with the highest fluoride concentrations included AZ13 (10.9 mg/L), AZ3 (7.32 mg/L), AZ12 (6.09 mg/L), AZ25 (5.75 mg/L), and AZ5 (5.66 mg/L), representing exceedances of 4 to 7 times the WHO drinking water guideline. Among irrigation wells (n = 8), two samples (25%) exceeded the WHO drinking water guideline, with the highest fluoride concentrations observed at AZ9 (4.15 mg/L), AZ14 (2.60 mg/L), and AZ17 (1.45 mg/L) (Fig. 3B). The generally lower fluoride levels in irrigation water (median 1.08 mg/L) compared to drinking water wells (median 4.40 mg/L) suggest that irrigation wells may tap less evolved groundwater with shorter residence times.

Nitrate concentrations

Nitrate concentrations showed the most extreme variability among the three contaminants of concern. Excluding sample AZ4 (3,433 mg/L), concentrations ranged from below detection to 184 mg/L (AZ21), with an overall median of 5.39 mg/L and mean of 19.4 mg/L. Drinking water wells (excluding AZ4) showed a median of 6.96 mg/L, while irrigation wells exhibited a similar median of 6.20 mg/L (Fig. 3C). Sample AZ4, a private drinking water well located at the Sultan’s residence in the urban center, exhibited extreme contamination indicative of severe point-source pollution from inadequate sanitation infrastructure, with concentrations 69 times the WHO guideline.

Among drinking water wells (n = 22), 3 samples (13.6%) exceeded the WHO guideline value of 50 mg/L: AZ4 (3,433 mg/L), AZ11 (99.8 mg/L), and AZ18 (66.7 mg/L) (Fig. 3C). Among irrigation wells (n = 8, including one sample below detection limit), one sample (12.5%) exceeded the WHO guideline: AZ21 (184 mg/L), located in the Teloua valley. This single exceedance suggests localized point-source contamination, possibly from nearby sanitation infrastructure or livestock activities, while the majority of sampled irrigation wells (86%) exhibited low nitrate levels below 15 mg/L. The generally similar nitrate levels between drinking and irrigation wells (medians 6.96 vs. 6.20 mg/L), but with different exceedance patterns, indicate that severe nitrate contamination is primarily associated with urban areas and inadequate sanitation infrastructure rather than diffuse agricultural practices.

Correlation analysis

Pearson correlation analysis (Fig. 4) suggests distinct geochemical controls on each contaminant. Uranium exhibited a moderate positive correlation with bicarbonate (R = 0.51), consistent with the formation of stable uranyl-carbonate complexes that enhance uranium solubility under near-neutral to alkaline conditions. Uranium showed weak correlations with other major ions.

Fig. 4.

Cross-correlation coefficients (Pearson R) between uranium, fluoride, nitrate, and selected hydrochemical parameters (n = 29, after one outlier removal, statistical significance indicated as **p < 0.01, ***p < 0.001).

Fluoride showed the strongest correlations with sodium (R = 0.80), pH (R = 0.55), and SO₄²⁻ (R = 0.66), and a negative correlation with magnesium (R = − 0.48). These associations point to fluoride mobilization under high-Na/Ca, alkaline conditions, a mechanism examined in detail in Sect.  5.2.

Across all 28 groundwater samples, nitrate showed a positive correlation with magnesium (R = 0.56), reflecting a combination of between-group differences and within-group processes. Irrigation wells exhibited systematically elevated Mg concentrations (8.7 vs. 3.8 mg/L, p = 0.033) and displayed a strong within-group NO₃-Mg correlation (R = 0.83, p = 0.020), consistent with co-contamination from agricultural sources (fertilizers, organic manure). In contrast, drinking water wells showed no significant NO₃-Mg correlation (r = 0.43, p = 0.055), indicating contamination from sanitation sources.

The cross-correlations between the three contaminants of primary concern were notably weak (Fig. 4). The correlation between uranium and fluoride was R = 0.13, between uranium and nitrate R = − 0.02, and between fluoride and nitrate R = − 0.20. None of these correlations approached statistical significance. The absence of meaningful correlations between U, F, and NO₃⁻ demonstrates that these contaminants derive from distinct and largely independent sources and mobilization mechanisms operating within the same aquifer system.

Hazard quotients and hazard indices

To assess the cumulative health risks for adults and children from multiple contaminants, we calculated the hazard quotients of uranium, fluoride and nitrate across the 22 drinking water wells (Fig. 5A,B) and assessed potential cumulative health risks due to co-exposure by assessing HI values.

Fig. 5.

Scatter plots showing the relationship between uranium hazard quotient (HQ(U)) and fluoride hazard quotient (HQ(F)) for 22 drinking water wells, color-coded by combined Hazard Index (HI) for (A) adults and (B) children. The HI represents the sum of individual HQs for uranium, fluoride, and nitrate. Gray dashed lines indicate the threshold of HQ = 1, above which non-carcinogenic health risks are considered significant.

The hazard quotients (HQ) and hazard indices (HI) for all individual drinking water wells are presented in Table 2.

Table 2.

Absolute concentrations of F, NO₃⁻ and U (< 0.01 = below detection limit) along with Hazard quotients (HQ) and hazard indices (HI) for adults and children.

Well	Type	F (mg/L)	NO3− (mg/L)	U (µg/L)	HQ(F)	HQ(NO3−)	HQ(U)	HI adult	HI child
AZ1	Private	5.53	0.98	209.1	2.63	0.00	1.99	4.63	10.80
AZ2	Private	4.89	< 0.01	53.7	2.33	0.00	0.51	2.84	6.63
AZ3	Public	7.32	< 0.01	1.3	3.49	0.00	0.01	3.50	8.16
AZ4	Private	0.95	3433	63.2	0.45	13.93	0.60	14.99	34.97
AZ5	Public	5.66	25.5	188.2	2.70	0.10	1.79	4.59	10.71
AZ6	Public	3.09	10.2	190.6	1.47	0.04	1.82	3.33	7.77
AZ7	Public	0.54	41.1	51.3	0.26	0.17	0.49	0.91	2.13
AZ11	Private	4.73	99.8	53.2	2.25	0.41	0.51	3.16	7.38
AZ12	Private	6.09	< 0.01	46.6	2.90	0.00	0.44	3.34	7.80
AZ13	Private	10.90	0.08	0.03	5.19	0.00	0.00	5.19	12.11
AZ18	Private	2.17	66.7	197.5	1.03	0.27	1.88	3.18	7.43
AZ19	Public	2.59	30.0	26.6	1.23	0.12	0.25	1.61	3.75
AZ20	Public	0.25	2.36	0.81	0.12	0.01	0.01	0.13	0.31
AZ22	Private	0.25	15.9	23.7	0.12	0.06	0.23	0.41	0.95
AZ23	Public	5.38	16.6	122.1	2.56	0.07	1.16	3.79	8.85
AZ24	Private	0.44	15.5	24.0	0.21	0.06	0.23	0.50	1.17
AZ25	Public	5.75	0.05	11.2	2.74	0.00	0.11	2.84	6.64
AZ26	Private	5.37	< 0.01	36.5	2.56	0.00	0.35	2.90	6.78
AZ27	Private	4.07	< 0.01	106.4	1.94	0.00	1.01	2.95	6.89
AZ28	Public	0.38	12.1	3.3	0.18	0.05	0.03	0.26	0.61
AZ29	Public	0.69	0.10	85.7	0.33	0.00	0.82	1.15	2.68
AZ30	Private	4.76	3.72	32.5	2.27	0.02	0.31	2.59	6.05

Reference doses: U = 0.003, F = 0.06, NO₃⁻ = 7.04 mg/kg/day. Bold values indicate HQ or HI > 1.0 (potential health concern).

Only five wells (AZ20, AZ28, AZ22, AZ24, and AZ7) showed HI ≤ 1.0 for adults, representing 22.7% of all drinking water wells, while not more than three wells (AZ20, AZ28, and AZ22) remained safe for children (13.6%). Notably, two of the three wells safe for both age groups are public supply wells (AZ20 and AZ28), with only one private well (AZ22) meeting safety criteria for children.

The combined Hazard Index (HI), representing the sum of individual hazard quotients for uranium, fluoride, and nitrate, exceeded the safety threshold of 1.0 in 77.3% of all assessed wells for adults (17/22) and 86.4% for children (19/22). HI values ranged from 0.14 to 14.99 for adults and from 0.32 to 34.97 for children. This disparity reflects children’s increased vulnerability due to higher water intake per body weight and lower body mass.

Wells with the highest combined health risks included AZ1 (HI = 4.63 adults, 10.80 children), AZ5 (HI = 4.59 adults, 10.71 children), and AZ6 (HI = 3.33 adults, 7.77 children), characterized by simultaneous elevation of both uranium (HQ(U) 1.79–1.99) and fluoride (HQ(F) 2.21–2.63). The extreme case of AZ4 exhibited the highest overall HI (14.99 adults, 34.97 children) driven predominantly by severe nitrate contamination (HQ(NO₃⁻) = 13.93 for adults), as indicated by the black circle marker in Fig. 5.

For adults, fluoride exceedances were most pronounced, affecting 60% of public wells compared to 75% of private wells (Fig. 6A). Uranium exceedances affected 30% of public wells and 25% of private wells. Nitrate exceedances were absent in public wells (0%) but affected 8% of private wells. The combined screening Hazard Index exceeded 1.0 in 70% of public wells and 83% of private wells, indicating that the majority of both well types pose cumulative health risks (Fig. 5A).

Fig. 6.

Comparison of health risk exceedances between public (n = 10, full coverage for Agadez) and private (n = 12, cross-sectional sample) drinking water wells for adults (A) and children (B). Bars show the percentage of wells with Hazard Quotient (HQ) for uranium, fluoride, or nitrate, or combined Hazard Index (HI), exceeding 1.0, the threshold indicating potential adverse health effects. Reference doses: U = 0.003, F = 0.06, NO₃⁻ = 7.04 mg/kg/day.

For children (Fig. 6B), exceedance rates increased across all categories due to their higher vulnerability. Fluoride exceedances are 60% in public wells and 83% in private wells. Uranium exceedances affected 50% of public wells and 58% of private wells. Nitrate exceedances remained at 0% for public wells and 8% for private wells. The combined HI exceeded safe levels (HI > 1) in 80% of public wells and 92% of private wells, with only one private well (AZ22) suitable for children’s consumption without treatment (Fig. 5B).

Discussion

Geochemical controls on uranium mobilization

The positive and statistically significant correlation between uranium and bicarbonate (R = 0.51, Fig. 4) provides evidence for carbonate complexation as the primary mechanism controlling uranium mobility in the Agadez groundwater system. Under oxidizing conditions typical of shallow to intermediate-depth groundwaters, uranium exists predominantly as uranyl (U⁶⁺) species³⁴. In carbonate-bearing waters, uranyl forms stable aqueous complexes with bicarbonate and carbonate ions, including UO₂CO₃⁰, UO₂(CO₃)₂²⁻, and UO₂(CO₃)₃⁴^−35,36. These complexes significantly enhance uranium solubility and inhibit adsorption onto mineral surfaces, allowing uranium to accumulate to high concentrations in groundwater³⁷.

Saturation index calculations provide further thermodynamic support for this interpretation. All groundwater samples are strongly undersaturated with respect to schoepite (UO₃·2 H₂O), the most common secondary uranium mineral in oxidizing environments, with SI values ranging from − 6.0 to − 2.2 (Fig. 7A).

Fig. 7.

Saturation index relationships for schoepite and calcite (A), and fluorite and calcite (B) in Agadez groundwater.

This persistent undersaturation indicates that no uranium mineral phase limits dissolved uranium concentrations; rather, uranium mobility is source-limited and controlled by the availability of uranium-bearing minerals in the aquifer matrix and the capacity of the groundwater to complex and transport uranium as uranyl-carbonate species.

The geological setting of the study area provides abundant sources for both uranium and the bicarbonate necessary for its mobilization. The Teloua sandstone formation contains feldspars and accessory minerals that host trace uranium of granitic origin (see also Sect. 2.2.). The Tim Mersoï Basin to the north of Agadez hosts major uranium ore deposits (e.g., Arlit and Akouta), and the regional geological formations exhibit elevated background uranium concentrations⁵. While active mining operations are located approximately 240 km northwest of Agadez, the same lithological units extend throughout the region, contributing to elevated uranium in groundwater through natural weathering and water-rock interaction.

Mechanisms of fluoride enrichment

The positive correlations between fluoride and pH (R = 0.55), and sodium (R = 0.80) (Fig. 4), are highly diagnostic of the well-established mechanisms controlling fluoride mobility in crystalline and sedimentary aquifers^38,10. Fluoride in groundwater originates primarily from weathering of fluoride-bearing minerals, including micas (biotite, muscovite), fluorapatite, and fluorite (CaF₂)³⁹. The release of fluoride from these minerals is enhanced under alkaline conditions, which promote hydrolysis of silicate minerals and dissolution of fluorapatite⁴⁰.

In the given case, fluorite solubility imposes a thermodynamic constraint on maximum fluoride concentrations: CaF₂ ⇌ Ca²⁺ + 2F⁻ (Ksp = 3.45 × 10⁻¹¹ at 25 °C). Saturation index calculations confirm this control: 60% of samples are near fluorite equilibrium (− 0.5 ≤ SI ≤ 0.5; Fig. 7B), reflecting calcite-fluorite coupling where Ca²⁺ removal by calcite precipitation promotes fluorite dissolution. In calcium-rich waters, fluorite precipitation removes fluoride from solution. Conversely, processes that deplete calcium – including cation exchange, precipitation of calcium carbonate, and evolution toward Na-dominated water types – allow fluoride to accumulate to higher concentrations⁴¹.

In the Agadez groundwater system, the hydrochemical evolution from Ca-HCO₃ type in shallow, rapidly recharged waters to Na-SO₄ type in more evolved waters creates conditions increasingly favorable for fluoride accumulation. The wells with highest fluoride concentrations (AZ13 at 10.9 mg/L, AZ3 at 7.3 mg/L) are characterized by high pH (8.1–8.3), strongly elevated Na/Ca molar ratios (52.8 and 32.2, respectively), and very low calcium (8.0–8.5 mg/L) and magnesium (0.13–0.14 mg/L) content, representing the hydrochemical endmember most conducive to fluoride enrichment. Calcium depletion is thus a necessary—though not sufficient—condition: while elevated fluoride is consistently associated with low Ca, calcium depletion alone does not guarantee high fluoride, as fluoride availability also depends on the presence of fluoride-bearing source minerals.

The presence of mica-bearing granites in the Air Massif and biotite as a minor constituent of the Teloua sandstone provides the primary source of fluoride to the groundwater system. Weathering of these silicate minerals releases fluoride along with sodium and silica, explaining the strong co-variation of fluoride with sodium concentrations observed in this study.

Sources and distribution of nitrate contamination

The distribution pattern of nitrate contamination, with highest concentrations in wells within the urban center and absence of systematic correlation with major ions in drinking water wells, clearly points to anthropogenic point sources^42,43. The absence of municipal wastewater collection and treatment infrastructure in Agadez means that domestic sewage is disposed through septic tanks and pit latrines, which can contaminate shallow groundwater through leakage and infiltration. In drinking water wells (n = 21), nitrate shows no significant correlation with major geochemical parameters, confirming that contamination follows localized point-source pathways. The extreme nitrate concentration at AZ4 (3433 mg/L) – located in the densely populated urban core – represents massive contamination from inadequate sanitation, likely direct infiltration of untreated wastewater.

In contrast, irrigation wells display elevated nitrate-magnesium correlation (r = 0.83, p = 0.020) and nitrate-calcium correlation (r = 0.77, p = 0.042), reflecting agricultural co-contamination from fertilizers and organic manure rather than urban sanitation sources. The observed correlation of NO₃⁻ to Mg across all samples (n = 29: r = 0.56, p < 0.01, Fig. 4) thus reflects between-group differences in land use and contamination pathways rather than a universal geochemical process.

The statistically insignificant correlation between nitrate and fluoride (R = − 0.20) and the negligible correlation between nitrate and uranium (R = − 0.02) confirm that nitrate contamination follows a pattern distinct from the geogenic contaminants. Nitrate enrichment preferentially affects the shallow, Ca-HCO₃ type groundwaters that receive more direct recharge from the surface, where anthropogenic nitrogen sources are concentrated.

Human health risk assessment of drinking water

The combined health risk assessment reveals that the majority of the population faces substantial chronic exposure to multiple contaminants. Hazard Index values for the 22 drinking water wells ranged from 0.14 to 14.99 for adults and from 0.32 to 34.97 for children (Fig. 5), where drinking water with HI exceeding a value of 1.0 is expected to have negative health effects under long-term exposure. The combined HI exceeded 1.0 in 77.3% of all wells for adults (17/22) and 86.4% for children (19/22) (Fig. 6), with 70% of public wells and 83% of private wells exceeding safe thresholds for adult exposure, and 80% of public wells and 92% of private wells for child exposure. The pattern of co-occurrence is particularly troubling as co-exposure to multiple contaminants represents very likely a more severe health threat than any single contaminant alone.

As groundwater serves as the sole drinking water source and this study encompasses all public drinking water wells from the urban area of Agadez, complemented by a cross-sectional sample of private wells, the given results reflect the principal quality of water consumed across the city. Thus, the public health implications for the 124,000 inhabitants are substantial. Given that approximately 70% of adults and 80% of children are exposed to drinking water from public infrastructure with combined hazard indices exceeding safe limits (HI > 1.0), and that private well users face even higher risks (83% for adults, 92% for children), the great majority of Agadez residents experiences chronic exposure to unsafe drinking water.

Furthermore, the health risk assessment based on drinking water consumption alone may underestimate total exposure, as some irrigation wells also exhibited elevated concentrations of fluoride, uranium and nitrate. The consumption of locally produced crops irrigated with contaminated groundwater represents an additional dietary exposure pathway that could increase the cumulative contaminant intake^14,15. Yet, unlike for drinking water, no human health-relevant thresholds for uranium and fluoride in irrigation water have been formally established in international guidelines. The FAO irrigation water quality guideline of 1.0 mg/L for fluoride⁴⁴ addresses primarily phytotoxicity and crop damage rather than direct human health risks through crop consumption, and no comparable guideline exists for uranium in irrigation water.

These findings underscore the urgent need for systematic water quality monitoring, public health interventions, and the implementation of appropriate measures to protect the population of Agadez from the cumulative health effects of uranium, fluoride, and nitrate exposure. Importantly, the weak inter-contaminant correlations indicate that uranium, fluoride, and nitrate originate from largely independent sources, which implies that no single intervention will address all three. Effective mitigation will therefore require contaminant-specific strategies: the geogenic contaminants uranium and fluoride call for source-based approaches, whereas the anthropogenic nitrate contamination requires improvements to local sanitation infrastructure and source protection. While advanced treatment technologies such as reverse osmosis or activated alumina are effective, their cost and maintenance requirements limit their feasibility in this resource-constrained setting. Simpler and more sustainable options are therefore likely to be more effective, including the preferential development of low-contaminant wells and the blending of water from contaminated and clean groundwater sources to dilute contaminant concentrations below guideline values. The few wells exhibiting consistently low concentrations of all three contaminants (e.g., AZ20 and AZ28) demonstrate that suitable source waters exist locally, although their limited number means that strategic well siting, protection of these resources, and expansion of centralized supply from low-risk aquifer zones would be required to make such an approach viable at the scale of the city.

Implications for Sahelian aquifer systems

The findings from Agadez have implications for groundwater quality across the Sahel. The geological and hydrogeochemical conditions promoting uranium or fluoride enrichment – weathering of crystalline and feldspathic sedimentary rocks, evolution toward alkaline, Na-rich, Ca-poor groundwater chemistry – are widespread throughout the Sahelian sedimentary basins and crystalline shield areas^5,38,10,3. Critically, the Precambrian crystalline basement rocks that underlie much of Sahelian West Africa – including granites, gneisses, and migmatites formed during the Pan-African orogeny – contain the same fluoride-bearing minerals (biotite, muscovite, amphiboles, fluorapatite) that contribute to groundwater fluoride enrichment in the Agadez region^18,2. Elevated groundwater fluoride concentrations have been documented in comparable geological settings across Burkina Faso⁴⁵, Nigeria^46,47, and Niger^2,11. Across these Sahelian regions, fluoride enrichment is consistently associated with crystalline (granitic) basement lithologies and groundwater evolution toward alkaline, sodium-dominated, bicarbonate-rich hydrochemical facies. Similarly, the sandstone formations of the Iullemeden Basin that host uranium mineralization near Agadez extend across Niger into Mali and Nigeria, with uranium-bearing geological units documented throughout the Tim Mersoï basin and related Paleozoic-Mesozoic sedimentary sequences^{5,48,24,25,22}.

While numerous studies have documented fluoride or nitrate contamination in West African aquifers⁴⁹, and others have examined uranium occurrence in arid and semi-arid groundwater systems⁵, no previous study has, to our knowledge, comprehensively assessed the co-occurrence of uranium, fluoride, and nitrate within the same groundwater systems in the Sahel. The present study therefore represents the first integrated, multi-contaminant assessment of uranium, fluoride, and nitrate exposure in groundwater used for drinking water in this region. Multi-contaminant groundwater quality problems are likely to be widespread in communities across Niger, Chad, Mali, Burkina Faso, and other Sahelian countries, but remain largely undocumented due to limited routine water quality monitoring capacity and a predominant focus on microbiological indicators and single contaminants^50,51.

The rapid population growth and urbanization occurring throughout the Sahel is simultaneously increasing water demand and intensifying anthropogenic contamination from inadequate sanitation. The combination of geogenic contaminants (U, F) with anthropogenic pollutants (NO₃⁻) creates a complex water quality challenge that cannot be addressed through single interventions. Addressing these complex, multi-contaminant groundwater challenges requires strengthened technical collaboration between local water authorities, research institutions, and international partners, particularly in drinking water development, quality monitoring, and capacity development⁵². Integrated approaches combining monitoring, source identification, locally appropriate treatment technologies, and sanitation improvements are urgently needed to support long-term groundwater sustainability and safe drinking water provision.

Conclusions

This study on groundwater quality in Agadez, Niger, has documented a triple threat to human health from naturally-occurring uranium and fluoride combined with anthropogenic nitrate contamination. The key findings are:

1. Uranium contamination in drinking water wells (n = 22) exceeds the WHO provisional guideline value of 30 µg/L in 63.6% of samples, with the highest drinking water concentration reaching 209.1 µg/L (AZ1) – nearly 7 times the guideline. Uranium mobility is controlled primarily by carbonate complexation.

2. Fluoride contamination is the most widespread concern, with 68.2% of drinking water wells (n = 22) exceeding the WHO guideline of 1.5 mg/L and maximum concentrations of 10.9 mg/L. Fluoride enrichment is controlled by weathering of fluoride-bearing silicate minerals under alkaline, Na-rich, Ca-poor conditions.

3. Nitrate contamination affects 13.6% of drinking water wells at levels exceeding the WHO guideline of 50 mg/L, with one extreme case (AZ4: 3,433 mg/L) representing severe point-source pollution from inadequate urban sanitation – notably, this well is actively used for drinking water purposes according to institutional records. Nitrate distribution follows anthropogenic source patterns and shows no significant correlation with geogenic contaminants.

4. Health risk assessment of drinking water wells indicates that the combined Hazard Index (HI) exceeded safe limits for 83% of private drinking water wells for adults and 92% for children (cross-sectional sample), compared to 70% of public supply wells for adults and 80% for children (full coverage), highlighting only a minor disparity in water quality between formal and informal water supply infrastructure.

5. Given that groundwater supplies the entire population of Agadez (approximately 124,000 inhabitants) with drinking water, a great majority are likely to face chronic health risks from combined uranium, fluoride, and/or nitrate exposure through drinking water consumption. Additionally, dietary exposure through locally produced crops irrigated with U and F contaminated water may further increase cumulative contaminant intake.

These findings highlight the urgent need for systematic groundwater quality monitoring and the implementation of appropriate treatment technologies for drinking water supplies in Agadez. To our knowledge, this study represents the first comprehensive assessment of the co-occurrence of uranium, fluoride, and nitrate contamination in Sahelian groundwater, addressing a critical knowledge gap for the region. The Precambrian crystalline basement rocks and Paleozoic–Mesozoic sedimentary formations that host fluoride- and uranium-bearing minerals in the Agadez region extend widely across the West African Sahel and associated sedimentary basins. Consequently, the hydrogeochemical conditions identified in this study are likely representative of those present in communities elsewhere in the Sahel, where similar geological settings and sanitation constraints may combine to produce multi-contaminant groundwater quality problems. Targeted assessments of uranium, fluoride, and nitrate in groundwater used for drinking water across the Sahel should therefore be prioritized to identify at-risk populations and inform interventions aimed at protecting public health.

Author contributions

Illias Alhassane: Conceptualization, Data curation, Formal analysis, Visualisation, writing-review and editing. Konstantin Scheihing: Conceptualization, Methodology, Investigation, writing-original draft. Manal Wannous: Methodology, project administration, validation, formal analyses, supervision, writing-review, editing. Bachir Tanimoun: Validation, formal analyses, writing-review, editing. Dennis Kraemer: Methodology, validation, resources, writing-review, editing. Robert Kringel: conceptualization, supervision, writing-review, editing.

Funding

This research was funded by the Federal Ministry for Economic Cooperation and Development (BMZ), Project No. BMZ PN 2019.2342.4.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.