Introduction

High Background Natural Radiation Areas (HBNRAs) are locations worldwide where natural radiation levels are significantly higher than average due to the presence of naturally occurring radioactive materials. Famous examples include Ramsar (Iran), Guarapari (Brazil), and Yangjiang (China)1,2. These areas are studied to understand the health effects of chronic low-dose radiation, as research hasn’t shown a significant rise in cancer rates, challenging the idea that any radiation exposure is harmful1,2,3. In certain HNBRAs, radiation levels are so elevated that they can exceed the occupational exposure limits set for radiation workers. These high doses are largely the result of unique geological characteristics specific to each region. For instance, Ramsar in Iran is influenced by hot springs rich in 226Ra, while Guarapari in Brazil lies on a belt of monazite sand1,2,4. In India, the coastal areas of Kerala and Orissa contain heavy mineral deposits with high monazite content, and Yangjiang in China has soils with a higher concentration of thorium than uranium1,2. These geological differences play a crucial role in determining the distribution and concentration of radionuclides, thereby shaping the environmental and health impacts unique to each HNBRA. Naturally elevated levels of ionizing radiation have been consistently observed in the Mamuju Regency of West Sulawesi, Indonesia5,6,7,8,9. A radiometric exploration has revealed that certain locations in Mamuju Regency, identified as sites of radioactive mineralization, exhibit average dose rates reaching 700 nSv/h, corresponding to an external terrestrial annual effective dose of approximately 5 mSv/y6. Further studies report that the total average annual effective dose in the region can reach 32 mSv/y, comprising an internal dose of 26.9 mSv/y and an external dose of 5 mSv/y8,9. The absorbed dose rates in these areas range from 50 nGy/h to 1109 nGy/h, with an average of 366 nGy/h10. In comparison, other well-documented HBNRAs, such as Yangjiang in China, report an average annual effective dose of approximately 2.1 mSv/y from internal exposure and 4.27 mSv/y from external exposure, yielding a total of about 3.5 mSv/y2,11. In Kerala, India, external annual effective doses range between 0.78 an In comparison, other well-documented HBNRAs, such as Yangjiang in China, report an average annual effective dose of approximately 2.1 mSv/y from internal exposure and 4.27 mSv/y from external exposure, yielding a total of about 3.5 mSv/yd 3.86 mSv/y, with absorbed doses between 650 and 3150 nGy/h (average 1925 ± 718 nGy/h)2,122,13. In Ramsar, Iran, public exposure in HBNRAs varies widely from 0.6 to 131 mSv/y, with an average of 6 mSv/y. Meanwhile, in Brazil’s HBNRAs, Poços de Caldas records the highest annual effective dose, averaging 13 mSv/y with a range of 6–16 mSv/y2,14. On a global scale, the typical annual radiation dose from internal sources is about 0.41 mSv, while the dose from external sources is much lower, at 0.07 mSv. The average outdoor radiation dose rate is approximately 59 nGy/h15. These findings suggest that Mamuju Regency qualifies as a HNBRA based on its significantly elevated radiation levels. The elevated concentrations of naturally occurring radioactive materials (NORM), including U, Th, and K, in the Mamuju region are closely associated with its unique geological history. This area is characterized by the presence of the Adang volcanic complex, whose past volcanic activity contributed to the formation of igneous and pyroclastic rocks enriched in radioactive elements. These geological processes led to the accumulation of uranium and thorium in concentrations that are atypical for Indonesia, making Mamuju the only known region in the country where such deposits are directly linked to volcanic origins6,16,17,18,19. Supergene enrichment is causing an increase in the concentration of radioactive elements like U, Th, and K in the surface soils of Mamuju6, which bears similarities to the processes observed in Poços de Caldas20.

As bedrock undergoes weathering in tropical environments, it transforms into iron-rich laterite through a process known as lateritization21,22. This transformation plays a key role in the secondary (supergene) enrichment of high field strength elements (HFSEs) such as U and Th and rare earth elements (REEs)19,23. These elements are often adsorbed onto clay minerals like halloysite, kaolinite, and smectite, which are common in lateritic soils19,22,23. The mobility of HFSEs varies, uranium is relatively soluble and can migrate with groundwater, whereas thorium is less mobile and typically binds to soil particles. Potassium, especially its radioactive isotope ⁴⁰K, contributes to natural background radiation and is readily taken up by plants24,25, although it can also be retained in clay-rich soils26. In Mamuju, the lateritization of bedrock has led to the formation of Fe₂O₃-rich soils23, where U, Th, and K are concentrated through these geochemical processes. This accumulation is influenced by a combination of environmental and geological factors, including the mineral composition of the parent rock, soil pH, and texture. U and Th are typically associated with specific igneous and sedimentary formations, and their concentrations in soil can increase over time due to prolonged weathering in tropical conditions6,16,17,18,22,23. The elevated levels of naturally occurring radioactive materials (NORM) observed in Mamuju’s surface soils present potential health risks to local populations9. Therefore, it is critical to investigate the mechanisms that govern radionuclide behavior in the environment. The primary objective of this study is to identify the environmental and geochemical factors of the soil surface that contribute to the elevated levels of natural background radiation in Mamuju Regency, West Sulawesi. The secondary objective is to characterize the radiological profile and understand the behavior of NORM in surface soils. This objective is accomplished by systematically examining key soil characteristics, including acidity, physical structure, and elemental content (via X-ray fluorescence), surface radioactivity (measured using HPGe gamma spectrometry), and ambient dose rates from field surveys. By integrating these measurements, this study provides a comprehensive understanding of NORM dynamics and the radiological characteristics of surface soils. The findings offer valuable insights into why Mamuju is classified as a High Natural Background Radiation Area (HNBRA).

Results

The pH and soil texture

The soil in Mamuju Regency is primarily derived from the weathering of volcanic rocks and is predominantly classified as laterite21,22 owing to its high aluminum oxide (Al₂O₃), ferric oxide (Fe₂O₃), and silica (SiO₂) content. Geochemical analysis using the XRF method confirmed these compositional characteristics22,27,28. The soil was found in acidic condition, with pH values ranging from 4.13 to 8.86 (mean: 5.84 ± 1.07), as measured by the standard electrometry methodVariability in soil texture was observed, with clay content ranging from 6% to 89%, sand content from 3% to 85%, and silt content from 8% to 60%. The distribution of these textures, classified according to the United States Department of Agriculture (USDA) system, is illustrated in Fig. 1. Table 1 summarizes the key physical characteristics of the soil samples, including pH values and textural data. These properties are essential for understanding the environmental conditions of the study area and their potential influence on the distribution of radioactive elements.

Fig. 1
figure 1

The USDA classification system for soil samples categorizes different soil textures based on their percentages.

Full size image

The XRF data

XRF analysis was conducted on 21 loose powder soil samples at BRIN’s Radiation Laboratory in Sleman, Yogyakarta, using the Omnian standard with Helium, to determine the key concentrations of radioactive elements (U, Th, K) and their major element oxides. Substantial variations in Th and U concentrations were identified, ranging from 139 to 3,520 ppm for Th and 43 to 578 ppm for U, while K fluctuated between 0.03% and 9%. The mean concentrations (236 ± 33 ppm U, 1,171 ± 195 ppm Th, and 2.55 ± 0.44% K) were found to be significantly higher than the global surface soil averages (2.8 ppm U, 10 ppm Th, and 2.4% K), indicating a pronounced geochemical anomaly in Mamuju soils. Statistical analysis confirmed these deviations (p < 0.05), and the distribution of U, Th, and K concentrations is summarized in Table 2.

The composition of Mamuju’s soils, heavily altered by weathering, shows significant variations in iron oxide (Fe₂O₃: 12%–60%), aluminum oxide (Al₂O₃: 4%–15%), and silicon oxide (14: 14%−38%). Trace element analysis corroborated this, with iron and aluminum as the main components. Furthermore, the relationship between soil acidity and the movement of U and Th (R² = 0.60 and 0.54) indicates that soil pH is a key factor in how these radionuclides disperse. Complete datasets, including elemental concentrations, are available in Supplementary Table S1. This comprehensive geochemical analysis underscores the significance of monitoring these elements to better understand the regional soil characteristics and their implications for natural radioactivity.

The chemical index of alteration and index of lateritization

Enhanced concentrations of U, Th, and K within Mamuju Regency are attributed to intense supergene enrichment and lateritization, processes driven by tropical/subtropical rainfall regions17,19,22,23. The extent of chemical weathering is assessed using a weathering index based on the ratios of specific chemical elements. In this study, the Chemical Index of Alteration (CIA)29 is applied to evaluate weathering intensity by comparing the ratios of immobile and mobile elements. Additionally, the degree of lateritization is quantified using the Index of Lateritization (IOL)21, which is determined as the ratio of SiO₂ to the combined total of Fe₂O₃ and Al₂O₃. A positive correlation between CIA and IOL was observed, with both indices exhibiting elevated values in regions characterized by high radioactive mineral potential. The detailed results of these calculations are provided in Supplementary Table S2.

Table 1 The soil acidity (pH) and textures from 21 sample locations.
Full size table
Table 2 The XRF data reveal concentrations of U, Th, and K, along with Th/U and Th/K ratios, complemented by precise soil radionuclide activity measurements obtained from HPGe gamma Spectrometry.
Full size table

Radionuclide activity

A detailed analysis using High-Purity Germanium (HPGe) gamma spectrometry was performed to determine the activity concentrations of radionuclides, specifically 226Ra, 232Th, and 40K, collected from 21 locations of the research site in Mamuju Regency. The measured average activities of 672 ± 16 Bq/kg for 226Ra, 1421 ± 30 Bq/kg for 232Th, and 619 ± 15 Bq/kg for 40K were recorded. These values considerably exceed the global average levels of 37 ± 4 Bq/kg for 238U, 33 ± 3 Bq/kg for 232Th, and 400 ± 24 Bq/kg for 40K. Uncertainties were determined by net peak area, gamma intensity, and sample weight30. These findings not only indicate the presence of potential radioactive mineral deposits6,16,31 and indicate a significant radiological risk to the population9. The detailed measurement data are presented in Table 2.

The environmental dose rate

A Canberra MiniRadiac Geiger-Müller Counter was used to measure environmental dose rates at 75 different locations, providing an effective assessment of local radiation levels. The measured average dose rate was 768 nSv/hr32, significantly higher than the global average for gamma radiation, which is 59 nSv/hr15. These findings indicate elevated radiation levels compared to the global background. Using QGIS, a 2D map was created to visualize the spatial distribution of gamma radiation dose rates for environmental assessment, not to locate uranium or thorium deposits32.

Discussion

pH and texture effect on U, Th, and K accumulation

The analysis was exclusively centered on the three elements U, Th, and K due to their radioactive nature and their role as the primary contributors to ambient radiation exposure. In this study, the impact of soil acidity levels was investigated to gain a deeper understanding of this phenomenon. As detailed in Table 1, the 21 soil samples were categorized into two distinct groups: those collected from prospective radioactive mineral areas (11 samples) and those from the surrounding vicinities (10 samples). An average soil pH of 6.02 was recorded across all samples, with the lowest pH of 4.13 recorded in Taloba Hamlet, Takandeang Village, and the highest at 8.86 in Uhayloppa, Labuanrano Village. Among the samples, 19 were shown to exhibit a pH less than 7, while 2 were found to be alkaline with a pH greater than 7. Soil pH exhibited a clear trend across the study area. Prospective radioactive mineral sites displayed acidic conditions, with pH values ranging from 4.13 to 6.28 (mean 5.30). Surrounding areas showed a slightly less acidic range of 5.43 to 8.86 (mean 6.66). Based on Havlin’s classification, acidity levels spanning from extremely acidic (4.5–5.0) to alkaline were found33,34.

The study reveals substantial concentrations of U, Th, and K. The concentrations range from 43 to 578 ppm for U, 139 to 3,520 ppm for Th, and 0.03% to 9% for K. Notably, the average concentrations are 236 ppm for U, 1,171 ppm for Th, and 2.55% (25,246 ppm) for K, with geometric means of 187 ppm, 878 ppm, and 1 ppm, respectively. These concentrations significantly exceed global averages of 2.8 ppm for U, 10 ppm for Th, and 2.4% (24,000 ppm) for K. The Th/U ratio, an indicator of radioisotope enrichment or depletion, ranges from 1.48 to 9.65, with an average of 5.0 ± 0.48, considerably higher than the Earth’s crust average of 315,35.

The findings underscore the strong influence of soil acidity (pH) on the concentrations of U, Th, and K36. High concentrations of U and Th correlate with increased soil acidity (low pH), demonstrating a clear negative correlation. In contrast, potassium concentrations rise in more alkaline soils (higher pH). This inverse relationship is effectively illustrated in Supplementary Figure S1 and the correlation analysis in Fig. 6. Soil texture also significantly impacts U and Th enrichment. Samples with higher clay content exhibit greater acidity. However, the high concentrations of U and Th are primarily linked to the properties of the clay itself19. Because of their extensive surface area and high cation exchange capacity, clay minerals are particularly effective at adsorbing and retaining these elements, a process enhanced by the weathering of the local U- and Th-rich rocks. This process is further enhanced by the weathering of local U- and Th-rich rocks, which promotes the formation of these clays19,37,38. XRD analysis of soil samples revealed dominant clay minerals, halloysite, kaolinite, and smectite, which have high adsorption capacities and are likely contributing to the accumulation of radionuclides in the region19.

The highest concentrations of U and Th were found in Taloba, where the extremely acidic soil (pH 4.13) and dominant clay texture created conditions favorable for their enrichment. This finding suggests a strong link between high soil acidity, clay content, and the retention of U and Th. Conversely, a very high concentration of K was detected in Hulu Ampalas, a prospective area for U and Th located away from populated regions6,7,39. The soil in Hulu Ampalas is classified as clayey sand, indicating a higher sand fraction that can enhance potassium (K) transport and leaching during weathering. Potassium is generally mobile in soil solution, particularly in coarse-textured soils with low cation-exchange capacity (CEC), due to reduced adsorption sites and increased hydraulic conductivity40,41,42. However, K mobility can be moderated by fixation on 2:1 clay minerals (e.g., illite and vermiculite) and its incorporation into primary minerals such as micas and feldspars43. Thus, K dynamics at this site reflect a balance between leaching in the sandy matrix and retention by the clay fraction. The measured soil pH of 6.09 falls within the slightly acidic range. The elevated K concentrations observed in Hulu Ampalas are likely derived from the parent material, which is notably rich in potassium (K). In contrast, lower K levels in other samples appear to be associated with more acidic pH values and may reflect soils formed from recent landslide deposits that have not yet experienced extensive weathering. The relation of pH, U, Th, and K with the clay, sand, and silt concentrations is described in Supplementary Table S4.

Weathering and lateritization

The enrichment of U, Th, and K in Mamuju Regency is closely tied to the weathering process. This process occurs as a result of intense weathering driven by tropical or subtropical rainfall, leading to secondary enrichment20. Chemical weathering involves the reaction of oxides with water, the decomposition of primary minerals in the parent rock, and the formation of secondary minerals. During this process, radioactive elements may also be mobilized and released into the soil. The release of U occurs through a redox reaction, and it subsequently infiltrates into the groundwater system44. Similar enrichment processes of U and Th have been observed in Pocos de Caldas, Brazil20, where the primary parent rock, phonolite, has undergone intense weathering under high rainfall and heat, leading to secondary enrichment in the soil and groundwater20. In Mamuju Regency, prolonged lateritization significantly alters the soil, which has a pH of approximately 6.5 and is enriched with Fe, Al, and Ti23. Silica and mobile elements such as K, Na, Ca, and Mg are leached by continuous rainfall. These elements are deposited in lower layers of the protolith, leaving the soil enriched in Fe, Si, and Al22. The weathering intensity was quantified using the CIA29, a metric derived from the ratio of immobile to mobile elements.

In this study, CIA values ranged from 15 to 98 (n = 21), with an average of 48 ± 5 (SD = 26), indicating considerable variability in weathering intensity among soil samples. Samples from locations with potential radioactive mineral deposits exhibited higher CIA values, with the highest recorded at Ahu Hill and the lowest at Hulu Ampalas (27–98, average 65 ± 8). The relatively low CIA values at Hulu Ampalas may be attributed to recent landslide events, which likely disrupted in-situ chemical weathering processes, leaving exposed fresh, unweathered rock and soil, effectively “resetting” the weathering process and resulting in a lower CIA value, even though the area is also a prospective site for uranium and thorium. In the HNBRA region, CIA values ranged from 67 to 99, with an average of 82 ± 7, suggesting moderate to strong weathering and enhanced chemical maturity of the soils45. These findings align with previous studies on tropical weathering trends in uranium-rich terrains45. Figure 2 illustrates the spatial distribution of weathering intensity across the study area.

Fig. 2
figure 2

The CIA ternary diagram clearly depicts the different levels of weathering present in the soil samples.

Full size image

The ternary diagram in Fig. 3 highlights a pronounced dominance of Fe2​O3​ (32.60%), indicating advanced lateritic weathering across all samples. Notably, despite the expected depletion of Si during lateritization, this study observed a concurrent enrichment of Si (10.57%) and Fe (22.79%), which may be attributed to secondary U-Th enrichment. A statistically significant negative correlation was observed between Fe2​O3​, Al2​O3​, U, and Th with soil pH (r = −0.58, −0.65, −0.77, and − 0.69, respectively; p < 0.05). However, no statistically significant linear relationship was found between SiO2​ and soil pH (r = −0.025; p > 0.05), suggesting that the behavior of silica in this context is not primarily driven by the same pH-dependent processes as the other elements21,46,47.

The Index of Lateritization (IOL), calculated using Eq. (3)21, serves as a quantitative measure of weathering intensity and varies based on the protolith composition. Figure 3 illustrates that all samples exhibit strong lateritization characteristics, with IOL values ranging from 33 to 81 (51 ± 19: n = 21). A strong positive correlation (r = 0.56, 0.75: p < 0.05) was observed between IOL and the enrichment of U and Th, with the highest IOL recorded in Botteng village and the lowest in Mekatta. These findings provide further evidence of lateritic processes influencing radionuclide distribution in high natural background radiation areas.

Fig. 3
figure 3

The IOL ternary diagram clearly depicts the different levels of lateritization present in the soil samples.

Full size image

The environmental dose rate and soil radionuclide activity

The presence of high concentrations of radioactive elements, specifically uranium (U: 42.5–578.1 ppm), thorium (Th: 138.8–3520 ppm), and potassium (K: 0.035–8.90%) in the soil samples, indicates significant exposure to natural radiation in Mamuju Regency. Mapping data confirms that Mamuju has the highest recorded natural radiation levels in Indonesia5, with potential radioactive mineral deposits identified in several locations6,7,39. This elevated radiation environment contributes to increased concentrations of radon (222Rn) and thoron (220Rn) in the air, which are well-documented risk factors for lung cancer and other radiation-related health effects9. Due to these factors, Mamuju is classified as a High Natural Background Radiation Area (HNBRA), a designation supported by its ambient gamma dose rates exceeding international safety benchmarks8,9,10,15.

The ambient dose equivalent H(10) was measured across 75 outdoor sites using a Canberra MiniRadiac Geiger-Muller Counter (GMC), yielding an average radiation dose rate of 774 ± 52 nSv/hour32. This value is significantly higher than the global average of 59 nSv/hour for environmental gamma radiation15. The measurements ranged from 108 to 2376 nSv/hour32, indicating significant spatial variability in radiation intensity.

A two-dimensional map, visually represents the spatial distribution of radiation exposure in the region, clearly shows that the highest radiation dose rates are concentrated in areas with significant uranium and thorium mineralization32. Even villages adjacent to these mineralized zones exhibit alarming dose rates, averaging 769 ± 453 nSv/h (with a range of 108 to 1059 nSv/h) among 75 measurement points32. These levels dramatically surpass Java Island’s typical background radiation of 52 ± 37 nSv/h48.

To quantitatively assess radionuclide activity, we analyzed 21 soil samples using HPGe Gamma Spectrometry (Canberra GC3518, FWHM = 1.77 keV at 1.33 MeV, calibrated against IAEA standards). The mean activity concentrations for 226Ra, 232Th, and 40K were 672 ± 16 Bq/kg, 1421 ± 30 Bq/kg, and 619 ± 15 Bq/kg, respectively—values significantly exceeding the global averages of 37 ± 4 Bq/kg, 33 ± 3 Bq/kg, and 400 ± 24 Bq/kg15. The highest activity levels were recorded in Taloba Hamlet, reaching 2114 ± 3 Bq/kg for 226Ra, 3869 ± 7 Bq/kg for 232Th, and 331 ± 4 Bq/kg for 40K. The distribution of radionuclide activity is presented in Supplementary Table S3.

The annual effective dose (AED)

The International Commission on Radiological Protection (ICRP) defines effective dose as a key measure for quantifying radiation exposure and its health risks. The global average for all natural background radiation, including cosmic, terrestrial, and internal sources, is about 2.4 mSv per year15.

This study measured the annual effective dose (AED) from outdoor terrestrial radiation in Mamuju Regency, finding significantly higher levels. Using a Canberra MiniRadiac detector, the average AED was 6.58 mSv, which is 2.83 times higher than the global average for terrestrial radiation of 2.4 mSv/y. The highest reading, 20.81 mSv/y, was found at Ahu Hill, a potential uranium-thorium site, while the lowest was 0.19 mSv/y at Labuanrano. The average dose in mineral-rich areas was 11.59 mSv/y, dropping to 4.7 mSv/y in surrounding regions. These levels often exceed the public exposure limit of 1 mSv/y and, in some cases, approach the 20 mSv/y occupational limit for radiation workers49. The distribution of the AED is illustrated in Fig. 4.

A comparison between AED values from a survey meter and those calculated from 21 soil samples (Fig. 5), revealed significant differences. The average AED from the soil samples was 1.46 mSv/y. This value is roughly half the global average of 2.4 mSv, but higher than the outdoor terrestrial AED average of 0.07 mSv15.The recorded value was approximately one-fifth of the estimate provided by the survey meter. The highest soil-based AED was 4.7 mSv/y in Taloba, and the lowest was 0.36 mSv/y in Mekatta.

These findings are consistent with earlier research indicating that the total natural background dose in Mamuju Regency can reach up to 32 mSv/yr8,9. Cumulative doses for some residents have been reported as high as 2.2 Sv, exceeding the exposure levels of some atomic bomb survivors9. Our study confirms that the soil surface radiation is a major contributor to the elevated outdoor terrestrial radiation levels in the region.

Fig. 4
figure 4

The annual effective dose from outdoor terrestrial radiation was determined using ambient dose equivalent (H*(10)) measurements collected with a survey meter.

Full size image
Fig. 5
figure 5

Comparison the AED values derived from soil radionuclide activity with those obtained from survey meter data.

Full size image

Physicochemical connection and statistical analysis

The Shapiro-Wilk test was performed at a 95% confidence level using OriginPro 2024 (Learning Edition) to assess data normality. The results (W = 0.89, p = 0.14) confirmed that the dataset deviated significantly from a normal distribution, justifying the use of the Spearman correlation method50.

As shown in Fig. 6, a Spearman’s rank correlation analysis highlighted the relationships between the physicochemical and radiological variables. Soil pH exhibited a positive correlation with K (ρ = 0.68, p < 0.05), ⁴⁰K (ρ = 0.72, p < 0.01), and sand content (ρ = 0.64, p < 0.01), indicating that higher pH values tend to be associated with elevated potassium concentrations and sand-dominated soils. In contrast, pH was negatively correlated with U (ρ = −0.79, p < 0.01), Th (ρ = −0.70, p < 0.05), ²²⁶Ra (ρ = −0.63, p < 0.01), ²³²Th (ρ = −0.69, p < 0.01), absorbed dose (ρ = −0.69, p < 0.01), and geochemical indices such as CIA (ρ = −0.75, p < 0.01) and IOL (ρ = −0.46, p < 0.01), suggesting that lower pH conditions favor the adsorption of radioactive elements.

A strong positive correlation was observed between absorbed dose from soil activity and dose rate from survey meter measurements (ρ = 0.65, p < 0.01), as well as geochemical indices like CIA (ρ = 0.85, p < 0.001) and IOL (ρ = 0.66, p < 0.01). This suggests that regions with high absorbed dose values exhibit correspondingly high radioactive element concentrations and elevated weathering indices. A stronger correlation was observed between CIA and the enrichment of U (ρ = 0.79, p < 0.001) and Th (ρ = 0.91, p < 0.001) compared to IOL. This indicates that the broader chemical weathering processes quantified by the CIA were more influential in driving uranium and thorium enrichment than the specific laterization processes represented by the IOL.

While there was a weak and statistically insignificant positive correlation between absorbed dose and silt content (ρ = 0.13,p > 0.01), suggesting no reliable relationship overall, a specific case was observed in Pasabu village. In this village, a sample with a dominant silt texture had a high measured radiation dose rate and a high absorbed dose. This finding contrasts with the strong positive correlations observed with sand content and potassium (ρ = 0.59 and ρ = 0.57, respectively), which are more significant factors influencing absorbed dose. Conversely, sand content exhibited a positive correlation with K (ρ = 0.59, p < 0.01) and ⁴⁰K (ρ = 0.58, p < 0.01), indicating that potassium (K) mineral concentrations are enriched in sandier soils.

The AED values derived from soil radioactivity and those measured using a survey meter show a statistically significant, strong positive correlation (ρ = 0.61, p < 0.001), indicating that both methods generally agree and validate each other’s results. A strong positive correlation (ρ = 0.69, p < 0.001) was observed between the calculated radiation dose from soil activity and the concentration of ²²⁶Ra, while a moderate positive correlation (ρ = 0.57, p < 0.01) was observed for ²³²Th. These relationships suggest that the survey meter reliably detected both isotopes, thereby providing statistical evidence that the two measurement methods were assessing the same underlying radiation source. All the detailed data are available in Supplementary Table S5.

Fig. 6
figure 6

A Spearman correlation of all Physicochemical variables.

Full size image

Multivariate statistic analysis

Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) were applied to uncover patterns, relationships, and groupings among variables influencing the enrichment of U, Th, and K in soil (Fig. 7). PCA simplifies complex datasets by converting correlated variables into a smaller set of independent principal components, enabling the identification of environmental factors, such as pH, CIA, soil texture, and dose rates, that are most strongly linked to U, Th, or K concentrations. HCA complements this by organizing variables or samples into clusters based on chemical, physical, and radiological similarities. Together, these methods help reveal underlying geochemical mechanisms, highlight variable correlations, and offer insights into the behavior of radioelements in the environment.

Principal Component Analysis (PCA) identified two main components, with PC-1 explaining 63.18% of the total variance and PC-2 accounting for 10.85%, indicating that PC-1 captures most of the variability in the dataset. PC-1 was primarily driven by radioactive elements (U and Th), their isotopes (²²⁶Ra and ²³²Th), radiation dose parameters (absorbed dose and AED from soil activity, dose rate and AED from the survey meter), and soil properties (CIA, IOL, and clay content). Prospective locations for radioactive mineral occurrences were generally positioned in the positive quadrant of PC-1, highlighting a strong association between geochemical characteristics and radioactivity-related parameters. Conversely, pH showed an inverse relationship, with higher values occurring in the negative quadrant of PC-1. Samples from Taloba, which plotted at extreme positive PC-1 values, displayed geochemical signatures typical of high-potential radioactive mineralization zones. Takandeang samples, despite being within the prospective zone, showed slightly elevated pH values, suggesting localized geochemical variation.

Fig. 7
figure 7

The linear combination relationship among all the Physicochemical variables is displayed in a biplot graph of multivariable correlation using the Principal Component Analysis (PCA) approach and dendogram of Hierarchical Cluster Analysis (HCA).

Full size image

PC-2, while accounting for less variance, revealed a distinct grouping of potassium-related variables (K, ⁴⁰K activity) with sand content and pH, indicating a strong positive correlation among them. This association was particularly evident in Hulu Ampalas samples, which plotted farthest from the origin in the negative quadrant of PC-2 and exhibited the highest values for these parameters.

Hierarchical Cluster Analysis (HCA) supported the PCA findings by forming three main clusters:

  1. 1.

    Cluster 1: pH, sand content, ⁴⁰K activity, and K, a cohesive group reflecting the potassium-rich nature of sandy soils, which appears to influence both K levels and soil pH.

  2. 2.

    Cluster 2: Radioactivity-related variables (U, Th, ²²⁶Ra, ²³²Th), radiation exposure parameters (absorbed dose and AED from soil activity, dose rate and AED from survey meter), and geochemical indicators (CIA, IOL, clay content), reflecting a shared geochemical process likely linked to lateritization and weathering. The close association between radiation exposure metrics and isotope concentrations underscores their common origin from the same radioactive sources in the soil.

  3. 3.

    Outlier (Silt): Silt content generally behaved independently of other variables. However, in Pasabu village, silt-rich samples contained unusually high U and Th concentrations and elevated radiation measurements. This suggests that, although silt is not generally a major carrier of radioactive elements in the region, specific local conditions can create a strong association.

Overall, the complementary use of PCA and HCA provided consistent evidence for two main geochemical–radiological groupings: (1) a potassium, sand, pH association, and (2) a radioactivity, clay, CIA, IOL association, with occasional localized deviations such as the Pasabu silt anomaly.

Conclusion

Soil acidity, texture, and weathering intensity exert a significant influence on uranium and thorium enrichment in the high natural background radiation areas of Mamuju Regency. Acidic soils (low pH), particularly those rich in clay minerals and exhibiting advanced lateritization, are associated with elevated U and Th concentrations, whereas potassium enrichment is more pronounced in alkaline soils. Geochemical weathering indices (CIA and IOL) indicate that zones of intense weathering and lateritization coincide with areas of high radionuclide activity.

Field surveys show that ambient gamma dose rates and annual effective doses (AED) substantially exceed global averages, frequently surpassing public safety limits and, in some locations, approaching occupational exposure thresholds. Measured activities of 226Ra, 232Th, and 40K in soils are several times higher than worldwide averages, with spatial clustering around U-Th mineralized zones.

PCA and HCA analyses consistently showed that soil radioactivity is primarily governed by specific geochemical factors. Two key relationships were identified:

  1. 1.

    A potassium–sand–pH association, where sandy, potassium-rich soils influence both potassium levels and soil pH.

  2. 2.

    A radioactivity–clay association, which links radioactive elements and radiation exposure to the presence of clay and the effects of weathering.

While radioactivity is generally tied to clay, some specific locations with high silt content were also found to have elevated radioactivity.

Materials and methods

The study adopts a hybrid methodology, integrating fieldwork and laboratory analysis. The fieldwork focuses on gathering qualitative data through direct observation, measurement, and examination of physical phenomena. This includes recording natural gamma radiation data with specialized instruments to map radiation dose rate patterns and collect soil samples. Laboratory analysis then builds on the fieldwork by examining the collected soil samples using a range of techniques in a controlled setting.

Study area

Several potential radioactive mineral deposits were identified in the Tapalang, Simboro, and Kalukku Districts6,31,32. Although a uranium deposit was reported near the crater of Mount Adang in Hulu Mamuju6,31, it could not be surveyed due to its remote location and difficult terrain. To establish control points, surveys also extended to nearby villages and areas exhibiting normal background radiation levels. Given the region’s complex landscape, including hills, valleys, forests, and riverbanks, the carborne survey method was deemed impractical. Instead, outdoor measurements were conducted using a systematic random grid approach. Vehicles were utilized only for accessing nearby locations, after which survey teams proceeded on foot to the designated points, as shown in Fig. 8.

Environmental dose measurement

The ambient dose equivalent (H*(10)) was measured by a team comprising two ITB students, one BRIN researcher, and two local workers, with additional personnel engaged as required by terrain and logistical demands. The equipment included a Mamuju sheet geological map, a Canberra Mini Radiac Geiger-Muller Survey meter, an Estwing E3 22P geological hammer, a Garmin eTrex 10 GPS, a mini shovel, a hoe, sample bags, ten 5-liter water jerrycans, and vehicles for transportation. The Canberra Mini Radiac used is a compact, lightweight, and durable radiation measurement device is designed for use by firefighters, HAZMAT teams, and paramedics32,51. It is equipped with an energy-compensated GM detector for gamma detection, offering ± 30% accuracy for dose rates between 100 µR/h (1 µSv/h) and 200 R/h (2.0 Sv/h), and ± 40% efficiency across the energy range of 60 keV to 1.5 MeV32,51. Prior to field deployment, the device was calibrated at the Safety Technology and Radiation Metrology Laboratory (LTKMR), Indonesia, using a substitution method with a standard 137Cs gamma radiation field, ensuring accurate dose rate readings in sievert (Sv) units52.

Measurements were taken at 1 m above the ground at 75 locations, ensuring consistency with international ambient dose assessment protocols (Fig. 8). Each measurement was repeated 3 times to reduce statistical error, and environmental conditions such as temperature, humidity, and potential shielding effects were considered. Supporting equipment, including a geological map, GPS unit, soil sampling tools, and transportation vehicles, facilitated efficient site navigation, positioning, and sample collection. These in situ radiation measurements complement the laboratory-based soil radioactivity analysis, providing a broader assessment of environmental radiation exposure in the study region32.

Soil sampling

21 Soil samples were collected from the surface layer at a depth of 0–20 cm, based on the consideration that this depth most significantly contributes to terrestrial gamma radiation as detected by the survey meter48. Sampling locations were selected based on previous radiometric exploration studies6,31,32, targeting areas identified as uranium-thorium (U-Th) deposits and their surrounding regions. At each location (Fig. 8), approximately 3 to 5 bulk soil samples were collected, with individual sample weights ranging from 3 to 5 kg. From these, a single sample was selected for laboratory analysis, specifically the one associated with the highest ambient dose rate recorded in situ using a survey meter. This approach assumes that the sample with the highest dose rate is representative of the area’s radiological characteristics.

Fig. 8
figure 8

The research site’s map in Mamuju Regency, West Sulawesi, Indonesia, displays ambient dose equivalent sampling points on the left and soil sampling points on the right.

Full size image

Laboratory work

Laboratory analysis of 21 soil samples was conducted to assess pH and texture at the Chemical Laboratory of the Agricultural Instrument Standard Testing Center, Agricultural Horticulture Instrument Standardization Agency, Lembang, West Bandung. These properties were analyzed to understand their potential influence on radionuclide distribution and retention. Soil pH was determined using the Electrometry method, calibrated with standard buffer solutions, yielding pH (H₂O) values, while soil texture was classified using the Gravimetry method, providing the percentage composition of clay, sand, and silt. Each location contributed 1 kg of carefully handled soil samples to prevent cross-contamination. The USDA soil texture classification was used.

Advanced testing, including XRF and HPGe Gamma Spectrometry, was performed using facilities provided by the National Research and Innovation Agency (BRIN) in collaboration with its researchers. XRF analysis, performed at the Yogyakarta Radiation Laboratory Facility using a PANalytical Epsilon4, employed a quantitative method to determine the concentrations of major and trace elements. Before analysis, 100–200 g of dried soil samples were sieved to a 100-mesh size to ensure uniformity.

The HPGe Gamma Spectrometry analysis was conducted at the Environmental Radiation Analysis Laboratory, BRIN, in Bandung, where calibration was performed using reference materials to ensure accuracy30. Approximately 500 g of homogenized soil samples were stored in 500 mL Marinelli beakers for over 30 days to reach secular equilibrium before measurement. The radionuclide activity concentrations of ²²⁶Ra (295 keV and 352 keV, 214Pb), 232Th (911 keV, 228Ac, and 583 keV, 208Ti) and ⁴⁰K (1460 keV) were determined using a Coaxial Canberra GX2518 with resolution 1.80 keV FWHM @ 1.33 MeV, multichannel analyzer (MCA) and Genie 2000 software (Canberra Industries), with a counting time of 17,000 s. The detection limits and uncertainties for these measurements were ensure for data reliability.

Estimation of CIA and IOL

The CIA and the IOL are geochemical indicators for assessing weathering intensity and lateritization processes. The CIA indicates the extent of chemical weathering in high-background radiation areas, while the IOL derived from SiO₂, Al₂O₃, and Fe₂O₃ contents, quantifies lateritization intensity in tropical soils. The CIA is calculated using the molar proportions of Al2O3, CaO, Na2O, and K2O29:

$$\:CIA=100\:\times\:\:\left[\frac{{Al}_{2}{O}_{3}}{{Al}_{2}{O}_{3}+CaO+{Na}_{2}O+{K}_{2}O}\right]$$
(1)

Similarly, the IOL is determined based on the proportions of SiO₂, Al₂O₃, and Fe₂O₃ to quantify lateritic weathering, using the formula21:

$$\:IOL=100\:\times\:\:\left[\raisebox{1ex}{$\left(Si{O}_{2}+{Fe}_{2}{O}_{3}\right)$}\!\left/\:\!\raisebox{-1ex}{$\left(Si{O}_{2}+{Al}_{2}{O}_{3+}{Fe}_{2}{O}_{3}\right)$}\right.\right]$$
(2)

Both indices were computed using XRF data, ensuring precise quantification of major element oxides.

Estimation of the external annual effective dose

To assess radiation exposure, the Annual Effective Dose (AED) was calculated based on soil radionuclide concentrations. This approach provides a detailed breakdown of dose contributions from specific isotopes, and it was used because indoor dose rate measurements were not available. The study therefore accounts only for the outdoor component of the annual effective dose.

We calculate the absorbed dose from the measured radionuclide activity data using the following Equation53.

$$\:D\left(\frac{nGy}{hr}\right)=0.462{C}_{Ra}+0.604{C}_{Th}+0.0417{C}_{K}$$
(3)

where, represent the activities of 226Ra, 232Th, and 40K.

First, the absorbed dose, D, is determine from Eq. (3). The AED was then derived from the calculated absorbed dose using the following relation:

$$\:{E}_{ext}\left(mSv\right)=D\bullet\:0.2\bullet\:0.7.\:8760\bullet\:{10}^{-6}$$
(4)

In this equation, D is the absorbed dose in nGy/h, 8760 is the annual exposure time in hours, 0.2 is the outdoor occupancy factor, and 0.7 is the dose conversion factor for adults from absorbed dose to effective dose.

This methodology, based on laboratory analysis of soil, provides specific radionuclide information and is not affected by short-term weather fluctuations. While real-time dose rates were also measured with a survey meter, the AED for this study was calculated from the detailed soil activity data.

Principal component analysis and hierarchical cluster analysis data preparation

There are approximately 16 datasets to be analyzed for their interrelationships. The majority of these datasets are expressed in the same concentration units, either in ppm or as a percentage. However, 7 of them differ; pH, CIA, and IOL are unitless, while absorbed dose and dose rate are measured in nGy/h and nSv/h. And the annual Effective Dose (AED) is measured in mSv. The remaining datasets use ppm or %. Soil radiation activity is expressed in Bq/kg but can be converted into concentration units (ppm or %), with conversion factors varying by element. Data normalization is necessary, one reason being that potassium (K), measured in %, has significantly higher values compared to uranium (U) and thorium (Th), as K is a major element in the Earth’s crust and naturally occurs in higher concentrations. Therefore, normalization must be applied before analysis. This process was conducted using the OriginLab Pro Learning Edition software, which offers a comprehensive Normalize tool with various normalization methods, ranging from scaling to a specific range, standardization via Z-scores, to division by statistical measures or reference-based normalization.

Statistical analysis, figure and map generation

All data visualization and statistical analyses, including the evaluation of significant relationships between variables, were conducted using OriginPro 2024 (Student Version, OriginLab Corp., USA). Spearman’s rank correlation analysis was performed to assess associations between variables, with correlation coefficients computed alongside a two-tailed significance test, applying a p-value threshold of 0.05. Additionally, 2D spatial maps were generated using the open-source Geographic Information System (GIS) software QGIS, version 3.22.3.