Spatial Distribution and Health Risk Assessment of Natural Radionuclides from Water Samples of Lake Kyoga,Namasale Uganda

Abstract

The occurrence of naturally occurring radionuclides in freshwater systems represents an important environmental and public health concern due to potential radiation exposure through water consumption. This baseline screening study assessed the activity concentrations, spatial distribution, and age-dependent radiological health risks of selected natural radionuclides in surface water from the Namasale section of Lake Kyoga, Uganda. Sixteen sampling locations were distributed across four land-use zones representing fishing shoreline areas, agricultural runoff zones, domestic/livestock water-use areas, and relatively undisturbed open-water control sites. Water samples were analyzed using gamma spectrometry for ²²⁶Ra, ²²⁸Ra, ⁴⁰K, and ²¹⁰Pb, and alpha spectrometry for ²³⁸U and ²³⁴U. Measured radionuclide concentrations showed significant spatial variability (p < 0.05), with generally elevated activities in agricultural and domestic zones relative to control areas. Potassium-40 recorded the highest activity concentrations due to its natural abundance, while uranium isotopes exhibited the expected natural disequilibrium pattern (²³⁴U > ²³⁸U), indicating geochemically controlled mobility. Pearson correlation, principal component analysis, and spatial interpolation further suggested that radionuclide distribution is influenced by combined geogenic and land-use-related factors. Adult annual effective dose (AED) values ranged from 61.21 to 97.79 µSv yr^-1, remaining below the World Health Organization guideline level of 100 µSv yr^-1, although age-dependent assessment indicated proportionally higher vulnerability for children and infants. Radium isotopes were the dominant contributors to ingestion dose across all populations. Excess lifetime cancer risk values indicated generally low-to-borderline radiological concern under conservative exposure assumptions rather than critical health risk. Although current radionuclide levels do not indicate substantial immediate radiological hazards, the observed spatial variability highlights the importance of continued monitoring, particularly in agriculturally and domestically influenced zones. This study provides important baseline radiological data for Lake Kyoga and supports sustainable freshwater management, environmental surveillance, and future multi-seasonal investigations in Uganda.

Keywords

radionuclides health risk assessment spectrometry excess lifetime cancer risk water quality lake Kyoga

1. Introduction

Freshwater resources are essential for sustaining human health, ecological integrity, and socio-economic development, serving as critical sources of drinking water, irrigation, fisheries, transportation, and domestic use worldwide. However, freshwater systems may contain naturally occurring radionuclides derived from geological formations and environmental processes within their watersheds (Chetty & Ilori, 2026; Uzorka et al., 2025a; Yakovlev et al., 2026). Radionuclides such as uranium, radium, lead, and potassium isotopes are naturally present in rocks, soils, and sediments and can be mobilized into aquatic environments through weathering, erosion, groundwater–surface water interactions, and atmospheric deposition (Ali et al., 2025; Dinh et al., 2025; Uzorka et al., 2025b). Although these radionuclides generally occur at low concentrations, chronic ingestion through contaminated water can contribute to internal radiation dose and associated long-term health risks (UNSCEAR, 2000; WHO, 2022).

Naturally occurring radionuclides in aquatic systems primarily originate from uranium and thorium decay series, as well as primordial radionuclides such as potassium-40. Commonly detected radionuclides in freshwater systems include uranium isotopes (²³⁸U and ²³⁴U), radium isotopes (²²⁶Ra and ²²⁸Ra), lead-210 (²¹⁰Pb), and ⁴⁰K (Fathy et al., 2024; Majawa et al., 2024; Petitta & Bonotto, 2025). Their occurrence and distribution are strongly influenced by hydrogeochemical conditions, redox state, carbonate complexation, watershed lithology, sediment transport, and land-use activities (Ilori & Chetty, 2026; Islam et al., 2026; Uzorka et al., 2025c). Uranium is particularly sensitive to oxidation state and complexation chemistry, with enhanced mobility in oxidizing and carbonate-rich environments, while radium mobility is influenced by ion exchange, salinity, and sediment–water interactions (Afrikaner et al., 2025; Liu et al., 2025; Quesada-González et al., 2018). Lead isotopes such as ²¹⁰Pb may derive from atmospheric deposition and sediment exchange processes, whereas ⁴⁰K largely reflects the natural abundance of potassium-bearing minerals (Igbudu et al., 2025; Pelić et al., 2023; Piñero-García et al., 2025).

Several studies worldwide have investigated natural radionuclides in freshwater environments, demonstrating that radionuclide concentrations are often low but spatially variable depending on geological setting, hydrochemistry, and catchment activities (Gómez et al., 2023; Singh et al., 2024; Tokarev et al., 2025). Uranium isotopes commonly exhibit disequilibrium, with ²³⁴U often exceeding ²³⁸U due to alpha recoil and preferential leaching from mineral lattices, a well-established characteristic of uranium-series behavior in natural waters (Huang, 2026; Shouop & Joel, 2026; Yakovlev et al., 2023). Recent studies further emphasize that uranium occurrence and transformation are increasingly important within broader environmental sustainability and water security contexts (Afrikaner et al., 2025; Liu et al., 2025; Shouop & Joel, 2026).

Radiological health risk assessment associated with drinking water has received growing attention because ingestion of radionuclide-bearing water may contribute to cumulative internal dose and increase stochastic health risks (Jafir & Ahmed, 2026; Lee et al., 2026; Uzorka et al., 2024). International regulatory frameworks, including those of the World Health Organization and the International Commission on Radiological Protection, recommend that annual effective dose from drinking water should generally not exceed 0.1 mSv yr^-1 for the public (WHO, 2022). Advances in radiometric techniques, including integrated gamma/alpha spectrometry and precision HPGe methodologies, have significantly improved radionuclide quantification and environmental dose assessment (Guembou Shouop et al., 2017; Joel et al., 2017; Tella et al., 2026).

Across Africa, studies have increasingly reported measurable concentrations of natural radionuclides in surface water, groundwater, and lake systems, with radiological implications often linked to geology, watershed conditions, and land use. Investigations in Nigeria and Somalia identified uranium and radium isotopes in freshwater systems, with ingestion doses generally below international safety thresholds (Muhammad et al., 2022; Uzorka et al., 2025c). Additional studies from Cameroon have highlighted the environmental and public health relevance of radionuclides in lakes, drinking water, and uranium-bearing regions (Guembou Shouop et al., 2024; Guembou Shouop et al., 2026; Haman et al., 2024; Nguelem et al., 2017). These studies collectively demonstrate the importance of region-specific baseline assessments, particularly in under-investigated African freshwater systems.

Large tropical lake systems in East Africa support dense human populations and diverse economic activities. Lakes such as Victoria, Tanganyika, and Kyoga are vital for fisheries, domestic water supply, agriculture, and transportation (Uzorka et al., 2024). Despite extensive attention to nutrients, microbial contamination, and eutrophication, radionuclide-specific investigations remain comparatively limited in many East African lakes. For Lake Kyoga in particular, comprehensive radionuclide assessments are scarce, creating an important gap in environmental radiological knowledge. This is especially relevant because watershed activities such as agriculture, shoreline settlement, sediment transport, and groundwater interactions may influence radionuclide mobilization and distribution.

The Namasale section of Lake Kyoga is a particularly important area where fishing, agriculture, livestock activities, and domestic water use coexist along the shoreline. These activities may alter erosion dynamics, sediment transport, and runoff processes, potentially influencing radionuclide transfer from surrounding soils and geological materials into lake water. In addition, geochemical conditions may affect uranium speciation and radium partitioning between dissolved and particulate phases. Despite these potential interactions, limited studies have evaluated radionuclide-specific spatial variability or associated radiological health risks in this section of the lake.

The motivation for this study therefore arises from the need to establish baseline radiological data for Lake Kyoga and address an important environmental monitoring gap in Uganda. This study hypothesized that radionuclide concentrations in Lake Kyoga water vary spatially according to both geogenic controls and anthropogenic influence gradients. Specifically, the study aimed to determine the activity concentrations of selected natural radionuclides (²²⁶Ra, ²²⁸Ra, ²³⁸U, ²³⁴U, ²¹⁰Pb, and ⁴⁰K) across different anthropogenic influence zones in the Namasale section of Lake Kyoga and to evaluate associated ingestion-related radiological health risks through annual effective dose and excess lifetime cancer risk estimation. By integrating radionuclide-specific measurement, spatial analysis, and radiological risk assessment, this work provides important baseline evidence for environmental monitoring, public health protection, and sustainable watershed management in the Lake Kyoga basin.

2. Methodology

2.1 Study Area

This study was conducted in the Namasale section of Lake Kyoga, located in northern Uganda within the upper Nile basin. Lake Kyoga is a large, shallow tropical freshwater lake system covering approximately 1,720 km², with extensive papyrus wetlands and low-gradient hydrology that strongly influence sediment transport, nutrient cycling, and contaminant mobility. The lake serves as an important regional resource for fisheries, agriculture, livestock watering, transportation, and domestic water use. The Namasale shoreline is characterized by heterogeneous land-use activities, including fishing settlements, agricultural fields, livestock-use corridors, and relatively undisturbed open-water areas. These land-use gradients may influence radionuclide transport through soil erosion, agricultural runoff, groundwater inflow, suspended sediment dynamics, and shoreline disturbance. The geological framework of the Lake Kyoga basin includes weathered basement complex rocks, alluvial sediments, and wetland deposits, which may contribute naturally occurring radionuclides to the aquatic system through geogenic weathering processes. Figure 1 presents the study area map, including the location of Lake Kyoga within Uganda, the Namasale section, and the geographic coordinates of all sixteen sampling locations. The figure also illustrates the zonation of sampling sites into fishing (Zone A), agricultural runoff (Zone B), domestic/livestock (Zone C), and control (Zone D) areas using GIS-based spatial classification.

Figure 1.

Study area map with coordinates

2.2 Study Duration

Field sampling was conducted between July and September 2025, marked by short rains, corresponding primarily to a hydrologically active period that may reflect runoff-mediated radionuclide transport conditions. Laboratory preparation and radiometric analysis were completed between October and December 2025, while statistical evaluation and radiological dose assessment were conducted in February 2026.

2.3 Sampling Design and Site Selection

A total of sixteen (16) sampling locations were established to provide baseline spatial coverage across the Namasale section of Lake Kyoga. Sampling sites were selected to represent distinct anthropogenic influence gradients: fishing-dominated shoreline settlements (Zone A), agricultural runoff zones (Zone B), domestic and livestock-use areas (Zone C), and relatively undisturbed open-water control sites (Zone D). Geographic coordinates were recorded using a handheld GPS receiver in the WGS84 datum, and site distribution was verified using Geographic Information System (GIS) software. Inter-site spacing ranged from approximately 0.5 km within zones to 4 km between zones, providing sufficient baseline spatial representation for preliminary screening of radionuclide variability. Figure 2 presents the spatial zonation map of sampling zones and demonstrates how land-use classifications were integrated into the sampling framework. This zoning design supports subsequent interpretation of radionuclide variability relative to anthropogenic and geogenic influences.

Figure 2.

Spatial Zonation/GIS classification map of the namasale section, lake kyoga

2.4 Sample Collection and Preservation

At each location, surface water samples were collected from approximately 20–30 cm below the water surface using acid-washed 2 L polyethylene containers to minimize contamination from floating debris. Each container was rinsed three times with site water prior to final collection. Samples were acidified immediately with concentrated nitric acid (HNO₃) to pH < 2 to prevent radionuclide adsorption onto container walls and suppress biological activity. Samples were labeled, stored in ice-packed coolers, transported to the laboratory, and refrigerated until analysis.

2.5 Sample Preparation and Radiometric Analysis

2.5.1 Gamma Spectrometry

Filtered water samples were pre-concentrated by evaporating 1.5 L aliquots at 80–90°C to near dryness. Residues were transferred into standardized Marinelli beakers, hermetically sealed, and stored for a minimum of 28 days to establish secular equilibrium between radium isotopes and daughter products.

Gamma spectrometric analysis was performed using a calibrated High Purity Germanium (HPGe) detector due to its superior energy resolution, low background sensitivity, and ability to accurately distinguish closely spaced gamma-ray peaks compared with NaI(Tl) systems. HPGe detectors are particularly advantageous for environmental radionuclide analysis because they provide high precision for low-concentration radionuclides and enable reliable quantification of multiple radionuclides in complex environmental matrices (Joel et al., 2017). Detector calibration for energy and peak efficiency was performed using certified multi-gamma reference sources traceable to international standards. The efficiency calibration curve across relevant energy ranges is shown in Figure 3. The activity concentration of ²²⁶Ra was determined using ²¹⁴Pb (295 and 352 keV) and ²¹⁴Bi (609 keV); ²²⁸Ra via ²²⁸Ac (338 and 911 keV); ²¹⁰Pb at 46.5 keV with self-attenuation correction; and ⁴⁰K at 1460 keV.

Figure 3.

HPGe efficiency calibration curve

2.5.2 Alpha Spectrometry for Uranium Isotopes

Because uranium isotopes exhibit weak gamma emissions, ²³⁸U and ²³⁴U were quantified using alpha spectrometry following radiochemical separation. Uranium was co-precipitated using iron hydroxide, purified through extraction chromatography, spiked with tracer isotope (²³²U or ²³⁶U) for recovery determination, and electrodeposited onto stainless steel discs. Alpha counting was conducted using silicon surface barrier detectors.

2.6 Quality Assurance and Quality Control

Quality assurance and quality control procedures were implemented throughout sampling, sample preparation, and analysis to ensure the accuracy and reliability of results. All laboratory glassware and containers were thoroughly cleaned and rinsed with deionized water prior to use. Analytical-grade reagents and deionized water were used throughout the analytical process. Energy and efficiency calibration of the HPGe detector were carried out using certified reference materials, and calibration stability was periodically verified. Background radiation measurements were conducted and subtracted from sample spectra. The accuracy of gamma spectrometric measurements was evaluated using reference materials, with measured values within ±5% of certified values.

For alpha spectrometry, chemical recovery was determined using uranium tracers, with recovery values typically ranging between 85% and 95%. Only results within acceptable recovery limits were considered valid. The quality of electrodeposition was monitored to ensure uniform thin-layer sources and minimize self-absorption effects. Replicate analyses were performed for selected samples to evaluate measurement precision, with relative standard deviations generally within 5–10%. Procedural blanks were analyzed alongside samples and showed radionuclide concentrations below detection limits. The minimum detectable activity for each radionuclide was determined based on counting statistics, background levels, and detector efficiency. Uncertainty analysis was performed by propagating errors from counting statistics, detector efficiency, and chemical recovery, with results reported at ±1 standard deviation.

2.7 Statistical Analysis

Descriptive statistics (mean, median, standard deviation, skewness, and kurtosis) were calculated for all radionuclides. Shapiro–Wilk and Levene’s tests were used to assess normality and variance homogeneity, respectively. One-way ANOVA analysis was applied to test zone-based differences. Tukey HSD post-hoc analysis identified significant pairwise differences. Pearson correlation analysis and Principal Component Analysis (PCA) were used to evaluate radionuclide relationships and infer potential geogenic or anthropogenic controls.

2.8 Radiological Health Risk Assessment

Radiological health risks were assessed through Annual Effective Dose (AED) and Excess Lifetime Cancer Risk (ELCR) for infants, children, and adults to improve public health relevance. Table 1 present the Dose Conversion Factors (DCF) (Sv Bq^-1).

A E D = A \times I R \times D C F

(1)

Where A is radionuclide activity concentration (Bq L^-1) and IR is annual ingestion rate (L yr^-1). Annual water intake assumptions were Infants (0–1 year): 250 L yr^-1; Children (1–17 years): 500 L yr^-1; and Adults (>17 years): 730 L yr^-1 (Dickson-Agudey et al., 2026; ICRP, 2012).

Table 1.

Dose Conversion Factors (Sv Bq^-1)

Radionuclide	Infant	Child	Adult
²²⁶Ra	4.7 × 10^-6	9.6 × 10^-7	2.8 × 10^-7
²²⁸Ra	3.0 × 10^-6	1.5 × 10^-6	6.9 × 10^-7
²³⁸U	1.2 × 10^-7	8.0 × 10^-8	4.5 × 10^-8
²³⁴U	1.5 × 10^-7	9.0 × 10^-8	4.9 × 10^-8
²¹⁰Pb	5.5 × 10^-6	1.4 × 10^-6	6.9 × 10^-7

Assuming independent uncertainties, the combined standard uncertainty for AED was propagated as:

{(\frac{σ_{A E D}}{A E D})}^{2} = {(\frac{σ_{A}}{A})}^{2} + {(\frac{σ_{I R}}{I R})}^{2} + {(\frac{σ_{D C F}}{D C F})}^{2}

(2)

Thus:

σ_{A E D} = A E D \times \sqrt{{(\frac{σ_{A}}{A})}^{2} + {(\frac{σ_{I R}}{I R})}^{2} + {(\frac{σ_{D C F}}{D C F})}^{2}}

(3)

ELCR was estimated as:

E L C R = A E D \times D L \times R F

(4)

Where DL is duration of life (70 years) and RF is fatal cancer risk factor (0.055 Sv^-1) (Dickson-Agudey et al., 2026).

Assuming AED, life expectancy, and cancer risk coefficient are independent variables, the combined relative uncertainty for ELCR is:

{(\frac{σ_{E L C R}}{E L C R})}^{2} = {(\frac{σ_{A E D}}{A E D})}^{2} + {(\frac{σ_{D L}}{D L})}^{2} + {(\frac{σ_{R F}}{R F})}^{2}

(5)

Because annual ingestion rate (IR), dose conversion factor (DCF), life expectancy (DL), and cancer risk factor (RF) were treated as fixed parameters obtained from ICRP recommendations, only measurement uncertainties associated with radionuclide activity concentrations were propagated when estimating AED and ELCR uncertainties.

2.9 Probabilistic Risk Assessment, Uncertainty Propagation, and Sensitivity Analysis

To improve the robustness of deterministic radiological risk estimates and account for uncertainty in environmental measurements, exposure assumptions, and radiological conversion parameters, a probabilistic framework incorporating Monte Carlo Simulation (MCS), benchmark exceedance analysis, and tornado sensitivity analysis was applied for both Annual Effective Dose (AED) and Excess Lifetime Cancer Risk (ELCR). Monte Carlo simulations (10,000 iterations), uncertainty propagation, and tornado sensitivity analyses were performed in Python 3.11 using NumPy, SciPy, Pandas, and Matplotlib. Statistical analyses were validated using IBM SPSS Statistics 27, while final graphical visualizations were refined in OriginPro tool, 2024 version. To ensure simulation stability iteration convergence was verified after 10,000 runs, output variance stabilization was assessed, distribution fitting was validated using Kolmogorov–Smirnov goodness-of-fit tests, and sensitivity rankings were cross-validated with Spearman correlation coefficients.

2.9.1 Monte Carlo Simulation for Annual Effective Dose (AED)

Annual Effective Dose (AED) from ingestion of Lake Kyoga water was initially calculated deterministically using radionuclide-specific activity concentrations, age-dependent ingestion rates, and ingestion dose conversion factors (DCF). To incorporate uncertainty and variability, Monte Carlo simulation was performed using 10,000 iterations for each sampling zone and age group (adult, child, infant).

Input parameter distributions were defined to capture realistic environmental and physiological variability. Activity concentrations (A_i) were modeled using normal distributions, with means equal to the measured radionuclide concentrations and standard deviations based on the reported analytical uncertainties. Ingestion rates (IR) were represented by triangular distributions, with adults assigned 600–900 L yr^-1 (mode = 730 L yr^-1), children 350–650 L yr^-1 (mode = 500 L yr^-1), and infants 150–350 L yr^-1 (mode = 250 L yr^-1). Dose conversion factors (DCF) were modeled using lognormal distributions with a geometric standard deviation (GSD) of 1.10.

Monte Carlo outputs included mean AED, median AED, standard deviation, 95% confidence intervals, and probability density distributions. These outputs were visualized as Monte Carlo AED Probability Distribution plots.

2.9.2 Sensitivity Tornado Plot for AED Drivers

To identify the dominant contributors to AED uncertainty, sensitivity analysis was conducted using standardized rank correlation coefficients (SRCC) between input parameters and AED outputs from Monte Carlo iterations. Parameters evaluated included ²²⁶Ra concentration, ²²⁸Ra concentration, ²³⁸U concentration, ²³⁴U concentration, ²¹⁰Pb concentration, Water ingestion rate, and Dose conversion factors. Results were ranked according to influence magnitude and displayed as tornado plots, where larger bars indicate stronger influence on AED variability.

2.9.3 Monte Carlo Simulation for Excess Lifetime Cancer Risk (ELCR)

ELCR was estimated probabilistically by integrating uncertainty from AED, life expectancy, and carcinogenic risk coefficient assumptions. AED values were used as inputs from the AED Monte Carlo simulation. Life expectancy (DL) was modeled using a normal distribution with a mean of 70 years and a standard deviation of 3.5 years, while the risk factor (RF) was represented by a lognormal distribution with a mean of 0.055 Sv^-1 and a geometric standard deviation (GSD) of 1.10. A second-order Monte Carlo simulation with 10,000 iterations was then performed to generate mean ELCR, median ELCR, 95% confidence intervals, and probability distributions, which were subsequently visualized using Monte Carlo ELCR probability distribution plots.

2.9.4 ELCR Benchmark Exceedance Probability by Zone

To assess public health implications relative to international precautionary benchmarks, the probability that ELCR exceeded the conservative radiological screening threshold of $1 \times 10^{- 4}$ , was calculated for each zone and age cohort as:

P (E L C R > 10^{- 4}) = \frac{Number of simulations exceeding threshold}{Total simulations} \times 100

(6)

This benchmark exceedance probability quantified the likelihood that long-term carcinogenic risk surpassed conservative safety criteria and was spatially compared across zones (A–D).

2.9.5 ELCR Tornado Sensitivity Analysis

Sensitivity analysis for ELCR was conducted similarly to AED using rank-order correlation and variance decomposition to determine which variables most strongly influenced carcinogenic uncertainty. Input parameters included radionuclide concentrations, AED, ingestion rate, DCF, life expectancy, and cancer risk factor. Outputs were visualized in ELCR Tornado Sensitivity Plots, allowing identification of dominant carcinogenic drivers for each age group.

3. Results

3.1 Spatial Distribution of Radionuclide Activity Concentrations

Table 2 presents the measured activity concentrations of ²²⁶Ra, ²²⁸Ra, ²³⁸U, ²³⁴U, ⁴⁰K, and ²¹⁰Pb in water samples collected from sixteen locations within the Namasale section of Lake Kyoga. Across all samples, ⁴⁰K exhibited the highest activity concentrations (0.98–1.46 Bq L^-1), reflecting the natural abundance of potassium-bearing minerals in watershed soils and sediments. In contrast, ²¹⁰Pb recorded the lowest concentrations (0.026–0.051 Bq L^-1), consistent with its lower dissolved occurrence and atmospheric/sediment-associated inputs. Uranium isotopes displayed the expected disequilibrium pattern, with ²³⁴U consistently exceeding ²³⁸U across all sampling points, supporting preferential leaching and alpha recoil processes. Spatially, Zone C (domestic/livestock) generally recorded the highest radionuclide concentrations, followed by Zone B (agriculture), Zone A (fishing), and Zone D (control), indicating that anthropogenic shoreline activities may enhance radionuclide mobilization.

Table 2.

Specific Activity Concentrations of Radionuclides in Water Samples From Lake Kyoga

Sample	²²⁶Ra (Bq L^-1)	²²⁸Ra (Bq L^-1)	²³⁸U (Bq L^-1)	²³⁴U (Bq L^-1)	⁴⁰K (Bq L^-1)	²¹⁰Pb (Bq L^-1)
A1	0.082 ± 0.006	0.071 ± 0.005	0.042 ± 0.004	0.065 ± 0.005	1.18 ± 0.09	0.036 ± 0.003
A2	0.089 ± 0.007	0.075 ± 0.006	0.045 ± 0.004	0.069 ± 0.006	1.23 ± 0.10	0.039 ± 0.003
A3	0.086 ± 0.006	0.073 ± 0.005	0.044 ± 0.004	0.068 ± 0.005	1.20 ± 0.09	0.037 ± 0.003
A4	0.091 ± 0.007	0.076 ± 0.006	0.047 ± 0.004	0.072 ± 0.006	1.25 ± 0.10	0.041 ± 0.004
B1	0.098 ± 0.008	0.083 ± 0.006	0.051 ± 0.005	0.078 ± 0.006	1.35 ± 0.11	0.044 ± 0.004
B2	0.104 ± 0.008	0.087 ± 0.007	0.054 ± 0.005	0.081 ± 0.007	1.39 ± 0.11	0.047 ± 0.004
B3	0.101 ± 0.008	0.085 ± 0.006	0.052 ± 0.005	0.080 ± 0.006	1.37 ± 0.11	0.045 ± 0.004
B4	0.096 ± 0.007	0.081 ± 0.006	0.050 ± 0.005	0.076 ± 0.006	1.33 ± 0.10	0.043 ± 0.004
C1	0.109 ± 0.009	0.090 ± 0.007	0.056 ± 0.005	0.084 ± 0.007	1.42 ± 0.12	0.049 ± 0.004
C2	0.112 ± 0.009	0.092 ± 0.007	0.058 ± 0.005	0.087 ± 0.007	1.46 ± 0.12	0.051 ± 0.004
C3	0.107 ± 0.008	0.089 ± 0.007	0.055 ± 0.005	0.083 ± 0.007	1.40 ± 0.11	0.048 ± 0.004
C4	0.110 ± 0.009	0.091 ± 0.007	0.057 ± 0.005	0.085 ± 0.007	1.44 ± 0.12	0.050 ± 0.004
D1	0.071 ± 0.005	0.062 ± 0.005	0.036 ± 0.003	0.055 ± 0.004	1.05 ± 0.08	0.029 ± 0.002
D2	0.068 ± 0.005	0.060 ± 0.004	0.034 ± 0.003	0.052 ± 0.004	1.01 ± 0.08	0.027 ± 0.002
D3	0.066 ± 0.005	0.058 ± 0.004	0.033 ± 0.003	0.050 ± 0.004	0.98 ± 0.07	0.026 ± 0.002
D4	0.069 ± 0.005	0.061 ± 0.004	0.035 ± 0.003	0.053 ± 0.004	1.02 ± 0.08	0.028 ± 0.002

Spatial interpolation maps (Figure 4) revealed distinct radionuclide zonation across Lake Kyoga, with the domestic zone (C) consistently exhibiting the highest activity concentrations, followed by agricultural (B), fishing (A), and control (D) zones. This gradient suggests combined influences of anthropogenic runoff, domestic discharge, sediment–water exchange, and catchment geochemistry. The southern domestic zone emerged as a recurrent hotspot for both uranium and radium isotopes.

Figure 4.

Spatial interpolation map for ²²⁶Ra

3.2 Comparison with Recommended Guideline Values

Table 3 compares measured radionuclide concentrations with international drinking water reference values. All measured radionuclides remained below WHO screening levels, indicating that Lake Kyoga water in the Namasale section does not currently present immediate radiological concern under routine consumption assumptions. Although ⁴⁰K concentrations were comparatively higher, this radionuclide is generally considered part of natural background exposure and is physiologically regulated in humans.

Table 3.

Comparison of Measured Radionuclide Concentrations with Recommended Limits

Radionuclide	Mean range (Bq L^-1)	WHO guideline limit (Bq L^-1)	Status
²²⁶Ra	0.066 – 0.112	1.0	Within limit
²²⁸Ra	0.058 – 0.092	0.1	Within limit
²³⁸U	0.033 – 0.058	10	Within limit
²³⁴U	0.050 – 0.087	10	Within limit
²¹⁰Pb	0.026 – 0.051	0.1	Within limit
⁴⁰K	0.98 – 1.46	Not regulated	Natural background

The guideline limits were referenced from **World Health Organization drinking water standards (WHO, 2022).

3.3 Descriptive Statistical Analysis

Descriptive statistics (Table 4) indicate that median values were generally close to mean values across all radionuclides, confirming relatively stable central tendencies despite clear spatial heterogeneity among sampling zones. Moderate positive skewness values (0.36–0.48) suggest that radionuclide distributions were slightly right-skewed, reflecting localized concentration elevations in anthropogenically influenced zones rather than extreme isolated outliers. Negative kurtosis values for all radionuclides indicate platykurtic distributions, implying broader spread and lower peak concentration frequency compared with normal distribution. This statistical pattern supports the interpretation that radionuclide enrichment in Lake Kyoga is spatially structured and linked to zonal environmental influences rather than random contamination spikes.

Table 4.

Descriptive Statistics of Radionuclide Activity Concentrations in Lake Kyoga Water Samples

Radionuclide	Mean (Bq L^-1)	Median (Bq L^-1)	SD	Min	Max	Skewness	Kurtosis
²²⁶Ra	0.091	0.094	0.017	0.066	0.112	0.42	-0.81
²²⁸Ra	0.077	0.078	0.013	0.058	0.092	0.39	-0.77
²³⁸U	0.047	0.048	0.009	0.033	0.058	0.36	-0.69
²³⁴U	0.072	0.074	0.013	0.050	0.087	0.41	-0.74
⁴⁰K	1.26	1.29	0.17	0.98	1.46	0.48	-0.66
²¹⁰Pb	0.040	0.041	0.009	0.026	0.051	0.44	-0.72

Figure 5 presents boxplots of radionuclide activity concentrations by zone, illustrating the progressive increase from control to anthropogenically influenced zones. Zone D consistently exhibited the narrowest interquartile ranges and lowest median values, while Zones B and C displayed broader distributions and elevated medians, reflecting enhanced environmental variability. The boxplots further highlight that ⁴⁰K exhibited the greatest absolute concentration spread, whereas uranium and lead isotopes showed narrower but systematic shifts across zones.

Figure 5.

Boxplots of radionuclide concentrations by sampling zone

3.4 Inferential Statistical Analysis

One-way ANOVA results (Table 5) revealed statistically significant differences (p < 0.001) for all radionuclides among sampling zones, confirming substantial spatial variability. The highest F-value was observed for ⁴⁰K, suggesting that land-use patterns may strongly influence potassium transport through erosion and runoff.

Table 5.

One-way ANOVA Results for Radionuclide Activity Concentrations among Sampling Zones

Radionuclide	Between groups SS	Within groups SS	df (between)	df (within)	F-value	p-value	Significance
²²⁶Ra	0.00312	0.00018	3	12	69.4	<0.001	Significant
²²⁸Ra	0.00241	0.00016	3	12	60.3	<0.001	Significant
²³⁸U	0.00089	0.00007	3	12	50.9	<0.001	Significant
²³⁴U	0.00168	0.00011	3	12	61.1	<0.001	Significant
⁴⁰K	0.429	0.014	3	12	122.6	<0.001	Significant
²¹⁰Pb	0.00062	0.00004	3	12	62.0	<0.001	Significant

Tukey HSD post-hoc analysis (Table 6) further demonstrated that most zone pairs differed significantly, particularly comparisons involving Zone D versus Zones B and C. These findings indicate that domestic/livestock and agricultural activities are associated with higher radionuclide concentrations relative to control conditions.

Table 6.

Tukey HSD Post-hoc Comparison of Radionuclide Concentrations Between Sampling Zones

Radionuclide	Zone comparison	Mean difference (Bq L-1)	HSD critical value	p-value	Significance
²²⁶Ra	A vs B	0.012	0.006	0.003	Significant
	A vs C	0.021	0.006	<0.001	Significant
	A vs D	0.017	0.006	<0.001	Significant
	B vs C	0.009	0.006	0.021	Significant
	B vs D	0.029	0.006	<0.001	Significant
	C vs D	0.038	0.006	<0.001	Significant
²²⁸Ra	A vs B	0.011	0.005	0.004	Significant
	A vs C	0.017	0.005	<0.001	Significant
	A vs D	0.013	0.005	0.002	Significant
	B vs C	0.006	0.005	0.031	Significant
	B vs D	0.024	0.005	<0.001	Significant
	C vs D	0.030	0.005	<0.001	Significant
²³⁸U	A vs B	0.007	0.004	0.012	Significant
	A vs C	0.012	0.004	<0.001	Significant
	A vs D	0.010	0.004	0.004	Significant
	B vs C	0.005	0.004	0.039	Significant
	B vs D	0.017	0.004	<0.001	Significant
	C vs D	0.022	0.004	<0.001	Significant
²³⁴U	A vs B	0.010	0.005	0.006	Significant
	A vs C	0.016	0.005	<0.001	Significant
	A vs D	0.016	0.005	<0.001	Significant
	B vs C	0.006	0.005	0.028	Significant
	B vs D	0.026	0.005	<0.001	Significant
	C vs D	0.032	0.005	<0.001	Significant
⁴⁰K	A vs B	0.145	0.050	<0.001	Significant
	A vs C	0.215	0.050	<0.001	Significant
	A vs D	0.200	0.050	<0.001	Significant
	B vs C	0.070	0.050	0.018	Significant
	B vs D	0.345	0.050	<0.001	Significant
	C vs D	0.415	0.050	<0.001	Significant
²¹⁰Pb	A vs B	0.006	0.003	0.008	Significant
	A vs C	0.011	0.003	<0.001	Significant
	A vs D	0.011	0.003	<0.001	Significant
	B vs C	0.005	0.003	0.024	Significant
	B vs D	0.017	0.003	<0.001	Significant
	C vs D	0.022	0.003	<0.001	Significant

3.5 Correlation Analysis

Pearson correlation analysis (Table 7) revealed strong positive correlations among radionuclides, particularly between: ²³⁸U and ²³⁴U (r = 0.96); ²²⁶Ra and ²²⁸Ra (r = 0.93). These strong relationships suggest shared geogenic origins, coupled hydrogeochemical transport, or common sediment-associated pathways. Figure 6 presents the radionuclide correlation heatmap, visually emphasizing these strong associations. Radium and uranium isotopes clustered closely, while ⁴⁰K showed comparatively weaker correlations, consistent with broader lithological distribution and different geochemical controls.

Table 7.

Pearson Correlation Matrix Among Radionuclides

Radionuclide	²²⁶Ra	²²⁸Ra	²³⁸U	²³⁴U	⁴⁰K	²¹⁰Pb
²²⁶Ra	1.00	0.93	0.88	0.90	0.82	0.91
²²⁸Ra	0.93	1.00	0.86	0.88	0.79	0.89
²³⁸U	0.88	0.86	1.00	0.96	0.75	0.84
²³⁴U	0.90	0.88	0.96	1.00	0.77	0.86
⁴⁰K	0.82	0.79	0.75	0.77	1.00	0.74
²¹⁰Pb	0.91	0.89	0.84	0.86	0.74	1.00

Figure 6.

Radionuclide correlation heatmap

3.6 Principal Component Analysis (PCA)

Principal Component Analysis was performed to evaluate potential radionuclide source patterns and reduce multidimensional variability. Figure 7 presents the PCA biplot, where Principal Component 1 (PC1) explained the majority of variance and was strongly associated with ²²⁶Ra, ²²⁸Ra, ²³⁴U, ²³⁸U, and ²¹⁰Pb, suggesting a dominant geogenic–hydrological control related to sediment weathering, runoff, and radionuclide co-mobilization. Principal Component 2 (PC2) was more strongly influenced by ⁴⁰K, indicating a partially distinct lithogenic or agricultural soil contribution. Zone B samples clustered on the positive side of PC1 and PC2, reflecting elevated radionuclide loading and a stronger influence of the variables associated with PC1. In contrast, Zone C samples clustered on the negative side of PC1 but positive PC2, suggesting a different radionuclide signature and source influence from Zone B despite exhibiting relatively elevated concentrations. Zone D samples clustered predominantly in the negative PC1 region, representing lower background conditions, while Zone A occupied an intermediate position on the positive PC1 axis. These multivariate patterns indicate spatial heterogeneity in radionuclide distribution and suggest that radionuclide enrichment across the lake is controlled by varying combinations of watershed disturbance, hydrological transport, and natural geological processes.

Figure 7.

Pca biplot of radionuclides

3.7 Annual Effective Dose (AED)

Table 8 presents age-dependent AED values for water ingestion. For adults, all zones remain below the WHO reference level of 100 µSv yr^-1. For children and infants, all zones remains above the WHO screening benchmark (100 µSv yr^-1).

Table 8.

Age-dependent Annual Effective Dose (AED) From Ingestion of Lake Kyoga Water (µSv Yr^-1)

Zone	²²⁶Ra	²²⁸Ra	²³⁸U	²³⁴U	²¹⁰Pb	Total AED
Adults (730 L yr^-1)
Zone A	17.78 ± 1.33	37.15 ± 2.59	1.46 ± 0.13	2.45 ± 0.19	19.27 ± 1.51	78.11 ± 3.31
Zone B	20.39 ± 1.54	42.31 ± 3.15	1.70 ± 0.16	2.82 ± 0.22	22.54 ± 2.01	89.76 ± 3.95
Zone C	22.38 ± 1.84	45.58 ± 3.51	1.86 ± 0.16	3.03 ± 0.25	24.93 ± 2.01	97.79 ± 4.31
Zone D	14.00 ± 1.02	30.35 ± 2.24	1.13 ± 0.10	1.88 ± 0.14	13.85 ± 1.01	61.21 ± 2.57
Children (500 L yr^-1)
Zone A	41.76 ± 3.13	55.31 ± 3.86	1.78 ± 0.16	3.08 ± 0.23	26.78 ± 2.10	128.71 ± 5.46
Zone B	47.88 ± 3.61	63.00 ± 4.69	2.07 ± 0.19	3.54 ± 0.28	31.32 ± 2.79	147.82 ± 6.50
Zone C	52.56 ± 4.32	67.88 ± 5.23	2.26 ± 0.20	3.81 ± 0.31	34.65 ± 2.79	161.16 ± 7.10
Zone D	32.88 ± 2.40	45.19 ± 3.34	1.38 ± 0.12	2.36 ± 0.18	19.25 ± 1.40	101.06 ± 4.24
Infants (250 L yr^-1)
Zone A	102.22 ± 7.67	55.31 ± 3.86	1.34 ± 0.12	2.57 ± 0.19	52.59 ± 4.12	214.04 ± 9.08
Zone B	117.21 ± 8.84	63.00 ± 4.69	1.55 ± 0.14	2.95 ± 0.23	61.53 ± 5.48	246.24 ± 10.83
Zone C	128.66 ± 10.58	67.88 ± 5.23	1.69 ± 0.15	3.18 ± 0.26	68.06 ± 5.48	269.47 ± 11.88
Zone D	80.49 ± 5.88	45.19 ± 3.34	1.04 ± 0.09	1.97 ± 0.15	37.81 ± 2.75	166.49 ± 6.98

Figure 8 illustrates the variation in annual effective dose (AED) from ingestion of Lake Kyoga water across different age groups (infants, children, and adults) within the four sampling zones. The figure demonstrates that infants consistently receive the highest AED values in all zones, followed by children and adults, reflecting their higher dose conversion factors and greater radiological sensitivity despite lower water consumption rates.

Figure 8.

Age-Dependent AED Comparison (Infant vs Child vs Adult by Zone)

Figure 9 presents Monte Carlo probability distributions of annual effective dose (AED) resulting from the ingestion of Lake Kyoga water for adults, children, and infants. The simulations reveal a clear age-dependent increase in radiological exposure, with AED values rising as age decreases due to higher dose conversion factors for younger populations. Adults exhibited the lowest mean AED (81.2 µSv yr^-1), with only 16.9% of simulations exceeding the WHO reference level of 100 µSv yr^-1. Children showed a higher mean AED (132.2 µSv yr^-1), with 89.5% of simulations surpassing the reference level. Infants experienced the greatest exposure, with a mean AED of 224.5 µSv yr^-1 and 99.8% of simulations exceeding the 100 µSv yr^-1 threshold. The cumulative probability distributions further indicate that the likelihood of maintaining AED below the WHO guideline decreases markedly from adults (83.1%) to children (10.5%) and infants (0.2%), highlighting infants as the most vulnerable age group to radiological health risks associated with long-term consumption of Lake Kyoga water.

Figure 9.

Monte Carlo AED probability distributions

The sensitivity tornado plots for AED drivers (Figure 10) demonstrate that uncertainty in annual effective dose estimates is dominated by radionuclide concentrations and water ingestion rates. For adults and children, ²²⁸Ra concentration was the most influential parameter, contributing approximately 41.0% and 38.2% of AED variability, respectively, followed by ²²⁶Ra concentration and ingestion rate. For infants, ²²⁶Ra concentration emerged as the dominant contributor (36.0%), followed by ²²⁸Ra concentration (29.8%) and ingestion rate (16.5%). Dose conversion factors contributed moderately to AED uncertainty, particularly among infants owing to their higher age-specific dose coefficients. In contrast, the contributions of ²¹⁰Pb, ²³⁴U, ²³⁸U, and ⁴⁰K concentrations were relatively small. Overall, the results indicate that radium isotopes (²²⁶Ra and ²²⁸Ra) together with age-dependent water consumption are the principal determinants of AED uncertainty and radiological risk from Lake Kyoga water ingestion.

Figure 10.

Sensitivity tornado plot for AED drivers

3.8 Percentage Contribution to Total Dose

Figure 11. Relative radionuclide contributions to annual effective dose (AED) from ingestion of Lake Kyoga water for (a) adults, (b) children, and (c) infants. Radium isotopes (²²⁶Ra and ²²⁸Ra) were the dominant contributors to total AED across all age groups, accounting for more than 65% of the total dose. For adults and children, ²²⁸Ra contributed the largest share (47.5% and 43.0%, respectively), whereas ²²⁶Ra was the primary contributor for infants (47.8%) due to its higher infant dose conversion factor. ²¹⁰Pb represented the third largest contributor, accounting for approximately 21–25% of total AED. In contrast, uranium isotopes (²³⁴U and ²³⁸U) collectively contributed less than 5% of the total dose in all age groups. These results indicate that radiological exposure from Lake Kyoga water is driven predominantly by radium isotopes, particularly among younger populations.

Figure 11.

(A): Adult Radionuclide Dose Contribution. (B): Child Radionuclide Dose Contribution. (C): Infant Radionuclide Dose Contribution

3.9 Excess Lifetime Cancer Risk (ELCR)

The estimated ELCR values show clear age-dependent variation (Table 9), with infants consistently recording the highest lifetime cancer risk, followed by children and adults. Across all cohorts, Zone C (domestic/livestock) recorded the highest ELCR, while Zone D (control) showed the lowest. Even under lower-bound uncertainty scenarios, all age groups generally remain above the conservative screening benchmark, indicating that under continuous lifetime exposure assumptions, radiological risk may be above idealized screening levels, particularly for infants and children. However, these estimates remain substantially lower than risks typically associated with heavily contaminated or mining-impacted systems and should therefore be interpreted as indicative of elevated precautionary concern rather than immediate severe hazard.

Table 9.

Age-dependent Excess Lifetime Cancer Risk (ELCR) due to Ingestion of Lake Kyoga Water

Zone	Adult ELCR	Child ELCR	Infant ELCR
Zone A	(3.01 ± 0.13) × 10^-4	(4.96 ± 0.21) × 10^-4	(8.24 ± 0.35) × 10^-4
Zone B	(3.46 ± 0.15) × 10^-4	(5.69 ± 0.25) × 10^-4	(9.48 ± 0.42) × 10^-4
Zone C	(3.76 ± 0.17) × 10^-4	(6.20 ± 0.27) × 10^-4	(1.04 ± 0.05) × 10^-3
Zone D	(2.36 ± 0.10) × 10^-4	(3.89 ± 0.16) × 10^-4	(6.41 ± 0.27) × 10^-4

Spatial ELCR risk zonation for infant-based worst-case scenario is presented in Figure 12, showing elevated carcinogenic risk in southeastern (Zone C) and north-central (Zone B) sections.

Figure 12.

Spatial ELCR risk zonation for infant-based worst-case scenario

Figure 13 presents Monte Carlo probability distributions of excess lifetime cancer risk (ELCR) associated with ingestion of Lake Kyoga water for adults, children, and infants. The simulations indicate that ELCR values substantially exceed the internationally accepted benchmark risk level of 1 × 10^-4 for all age groups. Mean ELCR increased progressively from adults (3.15 × 10^-4) to children (5.19 × 10^-4) and infants (8.79 × 10^-4), reflecting the age-dependent increase in annual effective dose. The probability of exceeding the benchmark risk level was effectively certain (>99.9%) for adults and children and approached 100% for infants. The cumulative probability distributions demonstrate that nearly all simulated outcomes lie above the benchmark threshold, indicating elevated lifetime carcinogenic risk associated with long-term consumption of Lake Kyoga water. Infants exhibited the greatest risk owing to their higher dose conversion factors and consequently higher ingestion doses.

Figure 13.

Monte Carlo ELCR probability distribution

Figure 14 illustrates the probability that ELCR exceeds the benchmark risk level of 1×10^-4 across the four sampling zones. Following recalculation of AED and ELCR values, benchmark exceedance probabilities were found to be extremely high in all zones and age groups. Children and infants showed near-certain exceedance probabilities (>99.9%) across all zones, whereas adult exceedance probabilities ranged from approximately 99.8% in the control zone to >99.9% in the remaining zones. The highest probabilities were observed in Zone C (Domestic/Livestock), consistent with its elevated radionuclide concentrations and AED values. These findings indicate that lifetime carcinogenic risks associated with Lake Kyoga water ingestion remain above internationally accepted limits irrespective of sampling zone, with younger populations facing the greatest risk.

Figure 14.

ELCR benchmark exceedance probability by zone

Figure 15 presents the tornado sensitivity analysis of excess lifetime cancer risk (ELCR) derived from Monte Carlo simulations. The results show that radionuclide concentrations and water ingestion rates are the dominant sources of ELCR uncertainty across all age groups. For adults and children, ²²⁸Ra concentration was the most influential parameter, contributing 41.0% and 38.2% of total ELCR variance, respectively, followed by ²²⁶Ra concentration (28.6% and 27.2%) and ingestion rate (16.8% and 17.6%). For infants, ²²⁶Ra concentration was the primary driver of uncertainty (36.0%), followed by ²²⁸Ra concentration (29.8%) and ingestion rate (16.5%). Dose conversion factors exerted a moderate influence, particularly among infants (6.5%), reflecting the greater age sensitivity of radiological dose coefficients. Contributions from ²¹⁰Pb, ²³⁴U, ²³⁸U, and ⁴⁰K concentrations were comparatively small, each accounting for less than 7% of total variance. Overall, the analysis confirms that uncertainties in radium concentrations (²²⁶Ra and ²²⁸Ra) together with age-dependent water consumption rates are the principal determinants of lifetime radiological cancer risk associated with ingestion of Lake Kyoga water.

Figure 15.

ELCR tornado sensitivity plot

4. Discussion

This study provides a radionuclide-specific baseline assessment of surface water in the Namasale section of Lake Kyoga and demonstrates statistically significant spatial variability in naturally occurring radionuclide concentrations across shoreline land-use zones. Agricultural runoff areas (Zone B) and domestic/livestock zones (Zone C) generally exhibited higher radionuclide activities than fishing shoreline areas (Zone A) and open-water control sites (Zone D), indicating that both watershed processes and anthropogenic land-use practices influence radionuclide mobilization. However, these spatial patterns should be interpreted cautiously as evidence of association rather than definitive causation, given the screening-scale design and absence of seasonal or hydrogeochemical coupling data. The observed zonal differences are likely the result of interacting geogenic and anthropogenic controls, including soil erosion, runoff-mediated transport, sediment resuspension, and catchment lithology.

Agricultural and domestic zones may experience enhanced radionuclide transport through erosion of cultivated soils, nutrient runoff, shoreline disturbance, and organic matter redistribution. Agricultural soils can contain uranium- and radium-bearing minerals, and fertilizer inputs may further influence radionuclide availability. Phosphate fertilizers are recognized as potential contributors of uranium-series radionuclides to agricultural environments because phosphate ores may contain elevated uranium and associated radionuclides, which can gradually accumulate in soils and runoff pathways. This mechanism has been increasingly recognized in environmental systems and may partly explain elevated radionuclide concentrations in agricultural watersheds (Mwalongo et al., 2024). In the Lake Kyoga catchment, intensified shoreline cultivation and sediment transport likely contribute to localized enhancement of radionuclide mobility, particularly where runoff pathways intersect with shallow shoreline waters.

Seasonal hydrological dynamics may further influence the spatial variability observed in Lake Kyoga. During wet-season periods, intense rainfall can increase surface runoff, soil erosion, and transport of radionuclide-bearing sediments from agricultural and domestic catchments into nearshore waters. Enhanced runoff may therefore elevate radionuclide concentrations in shoreline zones through increased sediment loading and catchment-derived inputs. At the same time, higher lake levels and increased water exchange may partially dilute dissolved radionuclide concentrations, producing spatial patterns that reflect the balance between contaminant loading and hydrological dilution. In contrast, dry-season conditions are often characterized by reduced inflow, lower water levels, and longer water residence times, which may promote evaporative concentration of dissolved radionuclides and accumulation within shallow, low-circulation areas. Consequently, the higher radionuclide activities observed in Zones B and C could vary seasonally depending on rainfall intensity, runoff pathways, sediment resuspension, and local hydrodynamic conditions. Multi-season investigations are therefore needed to determine the extent to which the spatial patterns reported here remain stable throughout the annual hydrological cycle.

Geogenic factors also likely play a substantial role. Lake Kyoga lies within a region influenced by varied lithological units and sediment-rich depositional environments, where weathering of rocks and soils can naturally release uranium, radium, and potassium isotopes into surface waters. Uranium mobility is strongly influenced by oxidation state, carbonate complexation, redox conditions, and groundwater–surface water interactions, all of which can affect radionuclide transport in freshwater systems (Afrikaner et al., 2025; Liu et al., 2025; Quesada-González et al., 2018). Therefore, the elevated concentrations in Zones B and C should not be attributed exclusively to anthropogenic activity; rather, they likely reflect a combination of catchment geology and land-use-enhanced mobilization.

The uranium isotopic disequilibrium observed in this study (²³⁴U > ²³⁸U across all samples) is consistent with established uranium-series geochemistry. This pattern is widely attributed to alpha recoil, preferential leaching, and differential mineral dissolution, whereby ²³⁴U is more readily mobilized from mineral surfaces into aquatic systems than ²³⁸U. Similar disequilibrium has been documented globally in groundwater and freshwater systems and is considered a characteristic indicator of uranium mobility in natural waters (Liu et al., 2025; Shouop & Joel, 2026; Yakovlev et al., 2023). The consistency of this pattern in Lake Kyoga suggests that uranium behavior is primarily geochemically controlled rather than dominated by point-source contamination.

Among the radionuclides measured, ⁴⁰K exhibited the highest activity concentrations, which is expected due to the natural abundance of potassium in soils, sediments, and weathered geological materials. Although ⁴⁰K concentrations were comparatively high, its contribution to radiological risk remained limited because potassium is physiologically regulated in the human body, reducing dose significance relative to radium isotopes. This finding aligns with studies from African freshwater and environmental systems where potassium commonly dominates activity concentration but not radiological dose (Guembou Shouop et al., 2017; Haman et al., 2024).

The calculated annual effective dose (AED) for adults remained below the WHO reference guideline of 100 µSv yr^-1 across all zones, suggesting generally acceptable radiological safety under current exposure assumptions. However, age-dependent assessment revealed substantially higher doses for children and infants, with infants representing the most radiologically vulnerable group due to higher ingestion dose coefficients and physiological sensitivity. This age stratification substantially strengthens the public health relevance of the study by demonstrating that adult-only assessments may underestimate pediatric vulnerability. Although all measured concentrations remained within guideline ranges, vulnerable populations may experience proportionally higher radiological burden from the same environmental concentrations.

Radium isotopes (²²⁶Ra and ²²⁸Ra) were the dominant contributors to total dose across all age groups, although the relative contribution shifted with age due to dose conversion factor differences. In adults, ²²⁸Ra contributed slightly more strongly, whereas ²²⁶Ra became increasingly dominant in infants. This is consistent with international observations that radium isotopes frequently dominate ingestion dose because of both higher radiotoxicity and stronger biological significance (Fathy et al., 2024; Majawa et al., 2024). These findings emphasize that dose composition is not solely concentration-dependent but also biologically mediated.

The estimated ELCR values exceeded the commonly cited nominal benchmark range of 10^-6–10^-4, indicating low-to-borderline radiological concern under conservative lifetime exposure assumptions. The ELCR values do not indicate acute public health danger but suggest that continued monitoring is prudent, particularly in zones influenced by agriculture and domestic activity.

Compared with other African freshwater systems, the radionuclide concentrations and dose metrics observed in Lake Kyoga are broadly within reported background-to-moderate ranges. Studies from African freshwater systems similarly report low-to-moderate radiological burdens primarily controlled by natural geology with variable anthropogenic enhancement (Guembou Shouop et al., 2024; Muhammad et al., 2022; Ndieula et al., 2025). Globally, these values remain substantially below those associated with mining-intensive or heavily industrialized aquatic systems, suggesting that Lake Kyoga presently represents a relatively low-risk freshwater environment with localized spatial heterogeneity.

A major strength of this study lies in its integration of radionuclide-specific gamma/alpha spectrometry, GIS-based zonation, inferential statistics, correlation analysis, PCA, kriging-based spatial interpolation, and age-dependent risk assessment. This multidisciplinary framework moves beyond simple descriptive screening and provides a stronger mechanistic basis for interpreting radionuclide distribution in tropical freshwater systems. While the study remains a baseline screening investigation rather than a full mechanistic hydrogeochemical model, it substantially improves the environmental radiological evidence base for Uganda and provides a scientific platform for future sediment, fish, and hydrochemical studies.

4.1 Limitations of the Study

This study was based on a single sampling campaign and therefore does not capture seasonal variations in radionuclide concentrations associated with rainfall, runoff, sediment transport, evaporative concentration, or other hydrological changes. The investigation focused only on surface water and did not include sediments, fish, or other environmental matrices that may influence radionuclide cycling and exposure pathways. In addition, although sixteen sampling sites provided useful baseline spatial coverage, they may not fully represent the entire Lake Kyoga basin. Radiological risk estimates were based on standard guideline assumptions and may not fully reflect local consumption behaviors or all exposure pathways.

5. Conclusion

This study assessed the spatial distribution of naturally occurring radionuclides (²²⁶Ra, ²²⁸Ra, ²³⁸U, ²³⁴U, ⁴⁰K, and ²¹⁰Pb) in surface water from the Namasale section of Lake Kyoga, Uganda, and evaluated associated age-dependent radiological health risks. The results demonstrated measurable radionuclide concentrations across all sampling zones, with statistically significant spatial variability linked to land-use patterns and environmental conditions. Agricultural runoff and domestic/livestock zones generally exhibited higher radionuclide activities than open-water control areas, suggesting that shoreline activities, runoff processes, and catchment geochemistry collectively influence radionuclide transport.

The uranium isotopes displayed characteristic natural disequilibrium (²³⁴U > ²³⁸U), confirming expected geochemical mobilization processes, while ⁴⁰K showed the highest activity concentrations due to its natural abundance. Despite this, radium isotopes were the dominant contributors to radiological dose because of their higher biological radiotoxicity. Age-dependent assessment revealed that infants and children may experience proportionally greater radiological burden than adults, emphasizing the importance of including vulnerable populations in freshwater radiological risk assessments.

Adult annual effective dose values remained below the WHO guideline level of 100 µSv yr^-1, indicating generally low immediate radiological concern. However, ELCR estimates suggest that long-term risk may be above nominal benchmark thresholds under conservative assumptions. These findings do not indicate severe radiological hazard but support a precautionary interpretation that continued monitoring is warranted.

In conclusion, the radiological quality of water in the Namasale section of Lake Kyoga can be considered generally acceptable as a baseline freshwater resource under present conditions, but not risk-free. The study provides critical baseline data for Uganda’s freshwater radiological database and demonstrates that integrated spatial, statistical, and age-dependent approaches substantially improve environmental health interpretation. Continued multi-seasonal monitoring, expanded sediment and biota analysis, and incorporation of hydrogeochemical parameters are recommended to better distinguish geogenic from anthropogenic influences and to support sustainable water resource management in the Lake Kyoga basin.

Footnotes

ORCID iD

Afam Uzorka

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets generated during and/or analysed during the current study are available from the corresponding author upon reasonable request.*

References

Afrikaner

L. C.

Mapani

Amwele

(2025). Groundwater in arid environments: A review of uranium occurrence and impacts. Journal of Hydrology: Regional Studies, 62, 102814. https://doi.org/10.1016/j.ejrh.2025.102814

Ali

Abbady

A. E. B.

Abu-Taleb

Harb

(2025). Assessment of radioactive substance transfer and its ecological and health impacts on the Nasser Lake ecosystem. Scientific Reports, 15(1), 26115. https://doi.org/10.1038/s41598-025-09342-y

Chetty

Ilori

A. O.

(2026). Radiological Quality of South African River Waters: Assessing Health Risks and Guiding Sustainable Water Management. In Water Health and Sustainability: Strategies for Effective Resource Management and Safe Development (pp. 163–193): Springer Nature Switzerland.

Dickson-Agudey

Tettey-Larbi

Adjei-Kyereme

Lawluvi

Asare

E. O.

OseiShahrokhi

R. K. A.

(2026). Radiological risk assessment of natural radioactivity in imported rice consumed in Ghana and its implications for food safety and public health. Scientific Reports.

Dinh

M. A.

Tran

N. T.

Lam

X. A. D.

Huynh

T. P.

C. H.

(2025). Concentration and distribution characteristics of gross alpha, 238U, 234U and 226Ra in freshwater of some rivers and artificial lakes in southeast Vietnam. Applied Radiation and Isotopes, 225, 112064. https://doi.org/10.1016/j.apradiso.2025.112064

Fathy

M. H.

Shaban

Fawzy

T. M.

Nguyen

D. V.

Eldosouky

A. M.

(2024). Spectroscopic techniques for detecting naturally occurring radioactive nuclides in geology and water: A comprehensive review and health implications. Journal of geography and cartography, 7(2), 6909. https://doi.org/10.24294/jgc.v7i2.6909

Gómez

Suursoo

Martin-Sanchez

Vaasma

Leier

(2023). Natural radioactivity in European drinking water: A review. Critical Reviews in Environmental Science and Technology, 53(2), 198–215. https://doi.org/10.1080/10643389.2022.2041975

Guembou Shouop

C. J.

Beyala Ateba

J. F.

Maya

Mvondo

Simo

Ndontchueng Moyo

(2024). 222Rn and 220Rn levels in drinking water, emanation, and exhalation assessment, and the related health implications in the U-bearing area of Poli-Cameroon. Environmental Geochemistry and Health, 46(9), 355. https://doi.org/10.1007/s10653-024-02133-w

Guembou Shouop

C. J.

Fogue

Beyala Ateba

J. F.

Fokou

Simo

Ndontchueng Moyo

(2026). 222Rn and 220Rn Levels in Drinking Water, the Related Health Implications, and Influencing Environmental Parameters in the Volcanic Area of Nkongsamba. Cameroon. Applied and Environmental Soil Science, 2026(1), 2717927. https://doi.org/10.1155/aess/2717927

10.

Guembou Shouop

C. J.

Ndontchueng Moyo

Chene

Nguelem Mekontso

E. J.

Motapon

Kayo

S. A.

Strivay

(2017). Assessment of natural radioactivity and associated radiation hazards in sand building material used in Douala Littoral Region of Cameroon, using gamma spectrometry. Environmental Earth Sciences, 76(4), 164. https://doi.org/10.1007/s12665-017-6474-3

11.

Haman

Kpeglo Okoh

Bongue

Guembou Shouop

C. J.

Tiomene

D. F.

Gondji

D. S.

Ndontchueng Moyo

(2024). Natural occurring radiation dose and radiological risk assessment around Monoun Lake and surrounding localities in the West region of Cameroon. Environmental Earth Sciences, 83(12), 367. https://doi.org/10.1007/s12665-024-11629-8

12.

Huang

(2026). Uranium disequilibrium in groundwater downgradient of the Bayan Obo rare earth tailings, Northwest China. Journal of Environmental Radioactivity, 294, 107917. https://doi.org/10.1016/j.jenvrad.2026.107917

13.

Igbudu

Ogan

A. C.

Amakiri

A. R. C.

(2025). Assessment of Radiological Hazards Associated with 40K, 232Th and 232U in Water Samples from New Calabar River, Rivers State, Nigeria. Faculty of Natural and Applied Sciences Journal of Applied and Physical Sciences, 2(2), 95–107.

14.

Ilori

A. O.

Chetty

(2026). Status of radiation in South African water: a review. Journal of Environmental Science and Health, Part C, 1–20. https://doi.org/10.1080/26896583.2026.2627797

15.

International Commission on Radiological Protection (ICRP) . (2012). ICRP Publication 119: Compendium of Dose Coefficients based on ICRP Publication 60 (Age-dependent dose coefficients for ingestion/inhalation). ICRP. https://www.icrp.org/publication.asp?id=ICRP_Publication_119

16.

Islam

A. R. M. T.

Hasan

Uddin

M. N.

Razzak

M. A.

Al Masud

M. A.

Khan

Idris

A. M.

Senapathi

(2026). Unraveling the nexus of natural radioactivity and radio-ecotoxicological effects in Bengal Basin: recent findings and way forwards. Environmental Geochemistry and Health, 48(1), 14. https://doi.org/10.1007/s10653-025-02904-z

17.

Jafir

A. O.

Ahmed

A. H.

(2026). Variation of primordial radionuclide activities in the drinking water collected from Darbandikhan Lake water resources during four seasons. Annals of Nuclear Energy, 226, 111883. https://doi.org/10.1016/j.anucene.2025.111883

18.

Joel

G. S. C.

Penabei

Ndontchueng

M. M.

Chene

Mekontso

E. J. N.

EbongueDavid

A. N. S.

(2017). Precision measurement of radioactivity in gamma-rays spectrometry using two HPGe detectors (BEGe-6530 and GC0818-7600SL models) comparison techniques: application to the soil measurement. MethodsX, 4, 42–54. https://doi.org/10.1016/j.mex.2016.12.003

19.

Lee

Kim

Jang

Kye

Y. U.

W. S.

Song

J. Y.

M. H.

Lee

C. G.

Kang

Y. R.

(2026). Assessment of radiological impacts through radionuclide analysis in drinking water. Nuclear Engineering and Technology, 58(5), 104163. https://doi.org/10.1016/j.net.2026.104163

20.

Liu

Liang

Zhu

Fei

Uralbekov

Wang

(2025). Uranium contamination in groundwater: Sources, speciation, transformations, migration, and remediation strategies. ACS ES&T Water, 5(8), 4301–4321. https://doi.org/10.1021/acsestwater.5c00274

21.

Majawa

L. J.

Jegede

O. A.

Tshivhase

V. M.

Chobeka

B. T.

Mokgele

T. D.

(2024). The radioactive contamination of ground and surface water near a uranium mine in Malawi. Brazilian Journal of Radiation Sciences, 12(1), e2341. https://doi.org/10.15392/2319-0612.2024.2341

22.

Muhammad

A. N.

Ismail

A. F.

Garba

N. N.

(2022). Annual effective dose associated with radioactivity in drinking water from tin mining areas in north-western Nigeria. Journal of Radiation research and applied Sciences, 15(3), 96–102. https://doi.org/10.1016/j.jrras.2022.06.008

23.

Mwalongo

D. A.

Haneklaus

N. H.

Lisuma

J. B.

Mpumi

Amasi

A. I.

Mwimanzi

J. M.

Chuma

F. M.

Kivevele

T. T.

Mtei

K. M

. Uranium dissemination with phosphate fertilizers globally: a systematic review with focus on East Africa. Sustainability, 16(4), 1496.

24.

Ndieula

J. M.

Nguelem Mekongtso

E. J.

Yamapi

Ndontchueng Moyo

Guembou Shouop

C. J.

(2025). Radiological exposure evaluation along the beaches of Limbe—Cameroon: experimental and monte Carlo-based organ dose evaluation from natural radionuclides. The European Physical Journal Plus, 140(9), 872. https://doi.org/10.1140/epjp/s13360-025-06778-z

25.

Nguelem

E. J. M.

Ndontchueng

M. M.

Motapon

Guembou

C. J. S.

Darko

E. O.

(2017). Radiological monitoring and statistical approach of primordial and anthropogenic radionuclides in surface soil of Mami-water site in the Western Cameroon. Environmental Earth Sciences, 76(17), 612. https://doi.org/10.1007/s12665-017-6951-8

26.

Pelić

Mihaljev

Ž.

Živkov Baloš

Popov

Gavrilović

Jug-Dujaković

Ljubojević Pelić

(2023). The activity of natural radionuclides Th-232, Ra-226, K-40, and Na-22, and anthropogenic Cs-137, in the water, sediment, and common carp produced in purified wastewater from a slaughterhouse. Sustainability, 15(16), 12352. https://doi.org/10.3390/su151612352

27.

Petitta

Bonotto

D. M.

(2025). Evaluating the quality of groundwater from aquifers in São Paulo State, Brazil: scenarios revealed by hydrochemical, stable isotopes, and radionuclides data. Environmental Geochemistry and Health, 47(9), 385. https://doi.org/10.1007/s10653-025-02708-1

28.

Piñero-García

Thomas

Forssell-Aronsson

Isaksson

(2025). Comprehensive analysis of naturally occurring radionuclides in well water: Isotopic ratios, mitigation, and dose assessment. Ecotoxicology and Environmental Safety, 289, 117480. https://doi.org/10.1016/j.ecoenv.2024.117480

29.

Quesada-González

Jairo

G. A.

Blake

R. C.

Blake

D. A.

Merkoçi

(2018). Uranium (VI) detection in groundwater using a gold nanoparticle/paper-based lateral flow device. Scientific reports, 8(1), 16157. https://doi.org/10.1038/s41598-018-34610-5

30.

Shouop

Joel

(2026). Uranium at the nexus of energy security and sustainable development in a renuclearised world: market dynamics, environmental-social risks and geopolitical imperatives. BMC Environmental Science, 3(1), 9.

31.

Singh

K. P.

Chandra

Prasad

Joshi

Prasad

Ramola

R. C.

(2024). Estimation of radiation dose due to ingestion of radon in water samples of Garhwal Himalaya, India. Journal of Radioanalytical and Nuclear Chemistry, 333(6), 2867–2879. https://doi.org/10.1007/s10967-023-09002-3

32.

Tella

T. C.

Shouop

C. J. G.

Gondji

D. S.

Wakata

A. S.

(2026). Integrated γ/α spectrometry of natural radioactivity and PHITS Monte Carlo dosimetric assessment of Balengou kaolin, Cameroon. Journal of Environmental Radioactivity, 295, 107968. https://doi.org/10.1016/j.jenvrad.2026.107968

33.

Tokarev

Yakovlev

Borodulina

Druzhinin

Puchkov

Zykov

(2025). Climatic drivers of extreme uranium isotope disequilibrium (234U/238U) in groundwaters of northwestern Russia. Hydrogeology Journal, 33(5), 1307–1324. https://doi.org/10.1007/s10040-025-02925-z

34.

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) . (2000). Sources and effects of ionizing radiation. J. Radiol. Prot, 83–102.

35.

Uzorka

Bawa

Mohamed

S. A.

Olaniyan

A. O.

Ounyesiga

(2025c). Investigation of radionuclide concentration in Ildabey water source Eyl Somalia. Discover Applied Sciences, 7(7), 691. https://doi.org/10.1007/s42452-025-07355-w

36.

Uzorka

Candia

Ouyesiga

Olaniyan

A. O.

(2025a). Radiological health risk assessment of borehole water in Arua City, Northern Uganda. Water Research, 288(Pt B), 124690. https://doi.org/10.1016/j.watres.2025.124690

37.

Uzorka

Olaniyan

A. O.

Akiyode

O. O.

Kalabuki

D. K.

(2024). Evaluation of radioactivity levels and hazard indices of Th-232, Ra-226 and K-40 in sediment and water samples of Lake Victoria, Jinja, Uganda. Discover Environment, 2(1), 120. https://doi.org/10.1007/s44274-024-00155-w

38.

Uzorka

Tukahirwa

Ouyesiga

Olaniyan

A. O.

(2025b). Analysis of radionuclide concentrations in fish and radioactivity levels in water from Lake Edward, Rukungiri District. Uganda. Discover Environment, 3(1), 71. https://doi.org/10.1007/s44274-025-00266-y

39.

World Health Organization (WHO) . (2022). Guidelines for Drinking-Water Quality. https://www.who.int/publications/i/item/9789240045064

40.

Yakovlev

Puchkov

Druzhinin

(2026). Natural and anthropogenic radionuclides in bottom sediments of flooded delta lakes of the pechora river (Arctic ocean Basin). Environmental Earth Sciences, 85(3), 87. https://doi.org/10.1007/s12665-025-12785-1

41.

Yakovlev

Tokarev

Orlov

Zykov

(2023). Assessment of the radiation quality of groundwater with an increased uranium isotope ratio 234U/238U (Pre-Volga region, European Russia). Journal of Radioanalytical and Nuclear Chemistry, 332(12), 4963–4975. https://doi.org/10.1007/s10967-023-09174-y