Influence of phosphate fertilizers on the radioactivity of agricultural soils and tobacco plants in Kenya, Tanzania, and Uganda

Three brands of NPK fertilizers that contain variable concentrations of natural radioactivity are commonly used in tobacco plantations in Kenya, Tanzania, and Uganda. Tobacco plants are known for hyper-accumulation of natural radionuclides, particularly 238U. This study investigated if the elevated radioactivity in phosphate fertilizers could enhance radioactivity in soils and tobacco plant leaves. The 232Th, 238U, and 40K radionuclide levels in NPK-fertilized soils and tobacco leaves were measured using gamma-ray spectroscopy. The research included a one-year reference experiment with tobacco growing in plots, a ten-year semi-controlled experiment in well-managed tobacco farms, and a field survey of radioactivity in soils and tobacco leaves at three traditional tobacco fields in Migori (Kenya), Urambo (Tanzania), and Kanungu (Uganda). The findings demonstrated that soils and tobacco leaves exposed to NPK fertilizers with increased radioactivity had activity concentrations of 232Th, 238U, and 40K that were considerably higher (at all sites) than in the control samples (with no use of NPK fertilizers). As the continued application of NPK fertilizers raises concentrations of 232Th, 238U, and 40K in agricultural soils, the study assessed radiological risks for humans from exposure to agricultural soils enriched with phosphate fertilizers, and it was found to be below the exposure limit of 1 mSvy-1 suggested by the International Commission on Radiological Protection (ICRP). However, tobacco consumers, both by snuffing and smoking, may face significant radiological risks, as the snuffing and smoking resulted in effective doses that were 2.41 to 6.53 and 1.14 to 2.45 times greater than the average yearly dose that the general public receives from inhalation of natural radionuclides (United Nations Scientific Committee on Atomic Radiations estimates). Furthermore, the results indicate that the lifetime excess cancer risk for tobacco snuffers and smokers ranged from 5 × 10-5 to 24.48 × 10-3 and 2.0 × 10-5 to 9.18 × 10-3, respectively. The influence of phosphorus-derived fertilizer containing relatively high natural radioactivity, potential human radiation exposure, and radiological risk due to gamma radionuclides is estimated and discussed. The results reveal that applying phosphate fertilizers enhances natural radioactivity in soil and is subsequently influenced by soil to tobacco plant uptake. Therefore, the study recommends that countries use fertilizers with lower radionuclide content to conserve soil quality and reduce gamma-emitting radionuclides in tobacco plants.


Introduction
The earth's crust contains various concentrations of the radioactive elements 238 U, 232 Th, and 40 K, which are all naturally occurring (Carvalho and Oliveira 2006;Gupta and Walther 2019). Phosphate ore, the raw material used to manufacture phosphate fertilizers, can contain naturally occurring radioactive elements such as uranium ( 238 U, T 1/2 = 4.47 × 10 9 a), thorium ( 232 Th, T 1/2 = 14.05 × 10 9 a), and potassium ( 40 K, T 1/2 = 1.25 × 10 9 a) in increased concentrations depending on the origin and type of rock and fertilizer manufacturing procedure, but often in concentrations that can be considered elevated (IAEA 2013(IAEA , 2018Haneklaus 2021;Van Dung et al. 2022). The application of phosphate fertilizers can thus be a major source of radionuclides and heavy metals added to agricultural soils (Haneklaus et al. 2017;De Souza Braz et al. 2021).

3
The amounts of naturally occurring radioactive materials (NORMs) added in a single fertilizer application may be considered insignificant compared to the volume of receiving soil and its naturally occurring radioactive elements. Nonetheless, it has been noted that the prolonged and intense use of phosphate fertilizers in tobacco crops led to a considerable buildup of 238 U, 232 Th, and 40 K in the soil (Khater 2012;Chauhan and Chauhan 2014;Bigalke et al. 2017Bigalke et al. , 2020De Souza Braz et al. 2021). Moreover, following fertilizer application, the additional radionuclides 238 U, 232 Th, and 40 K may move from the soil to other environmental compartments and represent a risk to human health through chemical and radiological exposure (Abed et al. 2022).
Natural radionuclides in fertilizers and agricultural soils have attracted the attention of researchers because of the potential radiation exposure to humans. For instance, Yamaguchi et al. (2009) examined the distribution of uranium in agricultural soil components following the long-term application of NPK fertilizer on rice paddies in Japan. They found that the uranium concentration was higher in NPK fertilized crops compared to the control sample. Takeda et al. (2006) reported that the uranium added with fertilizer application was concentrated in the top 35 cm of the soil stratum, with the soil organic compounds inf luencing uranium accumulation. In addition, the radioactive elements 238 U, 232 Th, and 40 K are long-lived and may remain in the agricultural soil, contaminate groundwater, and can even be taken up by plants (Papastefanou 2009;Gupta et al. 2020).
Uranium uptake by plants depends on the plant species, the amount of uranium in soils, and climatic conditions (Asaduzzaman et al. 2015;Tagami and Uchida 2020). Several studies have indicated that 238 U, 232 Th, and 40 K in the soil are not available for root uptake by most plant species. However, the tobacco plant (Nicotiana tabacum L.) has shown a distinct character as a hyper-accumulator of radionuclides from the soil compared to other plants, such as cereals and vegetables (Papastefanou 2009;Stojanovic et al. 2012;Kadhim and Ridha 2019;John et al. 2022).
Tobacco is a cash crop extensively grown in Kenya, Tanzania, and Uganda. Tanzania is a leading tobacco producer in the region, 3rd in Africa and 8th in the world (Ndomba 2018). Uganda is 2nd in East Africa, followed by Kenya, Rwanda, and Burundi (James 2019). In these countries, tobacco plantations intensively use nitrogen (N), phosphorous (P), and potassium (K) fertilizers from different brands (Lisuma et al. 2022b). Furthermore, Kenya, Tanzania, and Uganda are rural countries and more than 80% of the population depends on agriculture by cultivating various crops, including tobacco, for their livelihood (Lokuruka 2020(Lokuruka , 2021. The fact that large populations in these countries are directly involved in agricultural production means that a large fraction of the population could potentially be exposed to radioactivity beyond natural background levels through the intensive application of phosphate fertilizers. However, systematic studies on natural radionuclides associated with fertilizer use in tobacco production in Kenya, Tanzania, and Uganda have to the best of our knowledge, not been performed yet, although different studies have determined elevated concentrations of radionuclides, mainly 238 U in regionally produced and used mineral fertilizers in those countries (Makweba and Holm 1993;Banzi et al. 2000;Mwalongo et al. 2022).
It is worth mentioning that 210 Po and its precursor 210 Pb are primary sources of radiation exposure due to smoking (Carvalho and Oliveira 2006;Naina et al. 2010;Eke and Ishfaq 2021). Specifically, this study investigated the concentrations of 238 U, 232 Th, and 40 K in fertilizers, agricultural soils, and tobacco plant leaves in relationship with the application of NPK fertilizers. In addition, the study also used analytical data to assess the radiological risk and potential impact on the health of agriculture workers of tobacco plantations and consumers of this tobacco.

Experimental design and location of study areas
Three experiments were monitored, which included oneyear experiments, ten years of guided experimental farms monitored by the Tobacco Research Institute of Tanzania (TORITA), and smallholder farmer's practices' farms. The one-year experiment was a completely randomized design with four treatments. One district in each Kenya, Tanzania, and Uganda was selected for smallholder farmer's practice study.
The tobacco plant was chosen to assess soil-to-plant transfer of radionuclides based on its hyper-accumulation property of radionuclides, especially of uranium. Furthermore, tobacco is grown using NPK fertilizers rich in the macro-nutrients: nitrogen (N), phosphorous (P), and potassium (K) (NPK 10:18:24). The YaraMila Compound (YC) N 10 P 18 K 24 fertilizer contains three component nutrients: N, P, and K, while the YaraMila Blended (YB) N 10 P 18 K 24 fertilizer was blended mechanically to supply balanced nutrients of N, P, and K individually (Lisuma and Mbwambo 2022). In addition, the Golden Leaf Tobacco (GLT) N 10 P 18 K 24 fertilizer was blended from Minjingu organic Hyper-Phosphate (MoHP) rock containing significant concentrations of naturally occurring uranium (Mwalongo et al. 2022). Therefore, these mineral fertilizers can contain elevated concentrations of naturally occurring radionuclides associated with the phosphate ore processed for its P content.
Three studies were carried out. (1) A one year-control experiment with tobacco grown in plots was performed at the TORITA fields in western Tanzania in the 2021/2022 cropping season. (2) A long-term experiment was carried out in different tobacco fields in Tanzania under TORITA management for the previous ten years (from 2012/2013 to 2021/2022) and with annual tobacco cropping. (3) An assessment of radioactivity levels in traditional tobacco farms in the East Africa region was carried out in selected farms: the Urambo farm (Tanzania), Migori (Kenya), and Kanungu (Uganda).

The one year-controlled experiment
The experimental plots for the one-year controlled experiment were set up at the TORITA experimental farm in Tumbi (Tabora region) during the 2021/2022 cropping season. The tobacco plant Nicotiana tabacum L. (variety K326) was planted, and different brands of NPK fertilizers (NPK 10:18:24) were applied. The fertilizers used in this controlled experiment were YC, YB, and GLT. Tobacco growers used these NPK fertilizers during the experimental year. The first two fertilizers were imported from Sweden, while the last one was locally manufactured at the Minjingu Mines and Fertilizers Limited Company in Tanzania. After applying NPK fertilizers, calcium ammonium nitrate (CAN -27%) was applied to the experimental plots after two weeks as it is usually done in tobacco plantations. This fertilizer is based on synthetized ammonia and is practically radioactivity free. The three phosphate fertilizers (YC, YB, and GLT) were analyzed for gamma-emitting radionuclides at the gamma-ray spectrometry laboratory at the department of technical support and radiation protection service of the Tanzania Atomic Energy Commission (TAEC) Northern zone in Arusha.
The experimental plots were set up in agricultural soil that had not been cultivated for about ten years (a forest before) and received no influence from phosphate fertilization. The site was prepared by clearing the forest, and the soil was prepared using hand hoes. Tobacco seeds (variety K326) were sown in four seedbeds with four treatments (without fertilizer, fertilized with YC, YB, and GLT), each with 1 m × 1 m size after application of 1 kg of respective NPK fertilizer per treatment as is the standard procedure (Lisuma et al. 2022a).
Seedling seedbeds and experimental plots were prepared for one cropping season. Each experimental plot was replicated three times, thus obtaining a total of 12 plots (three plots per treatment). A control plot (no fertilizer) was also set up. The ridge and fallow field plots were 6 m × 6 m in size, with a 2 m spacing between blocks and a 1 m spacing within plots. A 3-meter-wide pathway was established around the entire experimental area to reduce potential outside influences. After 10-14 days, germinated tobacco seedlings were transplanted to new seedbeds and then re-transplanted into the experimental plots 60 days after sowing with an intra-row spacing of 0.5 m × 0.5 m.
Seven days after transplanting, the seedling's growth was boosted with application of the three fertilizer brands YC, YB, and GLT in the respective plots, each with three replications, as shown in Fig. 1. Twenty-one days after seedling transplanting, the CAN fertilizer (27% N) was applied at the rate of 30 grams per plant and agronomic practices were managed throughout the experiment. The tobacco leaves were sampled 12-15 weeks after topping.
The lower, middle, and upper-end leaves were sampled and mixed to get a composite sample for each sampling point for radionuclide measurement. In addition, a composite soil sample, combining soil from the ridge and fallow, was taken from each plot after reaping the leaves. The leaves and soil samples were packed in labeled plastic bags and transported to the laboratory for further preparation and analysis.

Ten-year semi-controlled experiment
The experimental sites used in this investigation were tobacco farms managed under the supervision of TORITA for the previous ten years (ten cropping seasons). One of the core functions of TORITA is to perform long term monitoring of the practices and performances of tobacco plantations in Tanzania. The tobacco farms selected had applied over the previous decade NPK (10:18:24) fertilizers, including NPK derived from YB and GLT, during the year 2021/2022. Therefore, farming practices for ten years duration were slightly different to the one year-control experiments. The tobacco farms investigated were located in 7 districts in Tanzania: Urambo (Tabora), Kahama (Shinyanga), Bukombe (Geita), Biharamulo (Kagera), Kakonko (Kigoma), Manyoni (Singida), and Namtumbo (Ruvuma) as shown in Fig. 2.

Investigation in selected tobacco farms in Kenya, Tanzania, and Uganda
The experiments in the three countries were carried out to assess the radioactivity levels in long-standing tobacco farms representative of the traditional tobacco growing districts/counties in Kenya (Migori), Tanzania (Urambo farm), and Uganda (Kanungu). Farmers' cultivation practices and NPK fertilizer application could not be controlled and might be slightly different from those followed in the one-year and ten-year experiments. These experimental sites at the three major tobacco producing sites were aimed at giving an actual picture of the radioactivity and radiation exposure in commercial tobacco farming. The districts/counties were chosen to represent leading tobacco producers. For instance, Urambo (Tanzania) accounts for over 50% of Tanzania's tobacco production (Lisuma et al. 2022b). Migori produces over 80% of tobacco in Kenya (James 2019), and in Uganda, the Kanungu district is responsible for about 20% of the countries tobacco production (Wanyonyi et al. 2020;Chune et al. 2022). Soil and tobacco leaf samples were collected from these farms and analyzed by gamma spectrometry at the TAEC in Arusha.

Sampling and sample preparation
Soil samples from a depth of 0-30 cm were collected using an open-end stainless-steel hand-held Auger corer (70 mm diameter) manufactured by Thermo-Fisher Scientific, Germany. Ten soil core samples were collected between the plants at 0-30 cm depth from each sampling site to make a layer of composite samples of about 500 g each. Larger plant debris and stones were removed, and soil samples were subsequently homogenized. Ten (10) tobacco leaves from the lower, middle, and upper portions were collected during the harvesting of the 2021/2022 cropping season. In the laboratory, the tobacco leaves were thoroughly washed with deionized water. Soil and tobacco samples were ovendried at 80 °C until a constant weight was reached. The dried soil and tobacco leaves were ground and sieved using a 200-micron mesh screen, transferred to metal canisters, and sealed airtight. The average weight of the sample materials in the canisters for gamma spectrometry was approximately

Radionuclides analysis by gamma spectrometry
After 30 days of storage to allow for the formation of secular radioactive equilibrium of 238 U and 232 Th with their respective short-lived progenies, while 40 K was measured from the 1640 keV distinct peak, the samples were counted for 36,000 s on a large volume high purity germanium (HPGe) gamma spectrometer (manufactured by ORTEC), inside lead shielding and connected to a multichannel analyzer. The Gamma Vision ® software was used in spectrum analysis (ORTEC 2020). The energy and efficiency calibration of the TAEC gamma spectrometry system was performed for lower and higher energies using multi-nuclide calibration sources type CBSS 2, Certificate No. 1035-Se-40202-17, Serial No. 270217-1621040, trace-able to the Czech Metrology Institute. The average concentrations of radionuclides 212 Pb (238.6 keV) and 228 Ac (911.1 keV) were used to calculate the concentration of 232 Th in the sample, while the peak gamma energies of radionuclides 214 Bi (764.8 keV) were used to calculate the concentration of 238 U. The gamma lines 609.3 keV and 1120 keV for the 238 U series, and 583.2 keV, 727.3 keV, and 795 keV for the 232 Th series were removed from the estimation of radionuclide activity concentration because they are impacted by coincidence summation (Newman et al. 2008;Polouckova 2021). The activity concentration for each radionuclide of interest was calculated using Eq. 1 (Kovacs et al. 2017): where A is the activity per unit mass of dry weight of the sample in Bq kg −1 , N is the net counts per second (cps), is a gamma-line emission probability of a particular radionuclide, ε is the gamma line emission intensity (%), and (kg) is the weight of the sample.

Analytical quality control
Using the IAEA-375 Certified Reference Material (CRM) with soil matrix, the calibration of the spectrometry equipment was evaluated for accuracy and reliability. The outcomes are displayed in Table 1.
The relative bias between the IAEA certified reference values and measured results were calculated using Eq. 2 The results indicate that the overall bias was within ±8%, which indicate that the accuracy of the measurement method was suitable for the purpose of this research (Shakhashiro et al. 2008).

Physical-chemical characteristics of soils
The physical-chemical characteristics of the soil samples, namely, cation exchange capacity (CEC), organic matter content (OM), soil pH, and soil textures (sand, silt, and clay), were determined to the depth of 0-30 cm using standard scientific methods described by Moberg (2001).

Calculation of radiological parameters
It was hypothesized that radionuclides present in soil and tobacco leaves could be a radiological hazard to farm workers and members of the public. Furthermore, tobacco consumers may ingest and inhale the radionuclides contained in the tobacco leaves (IAEA 2013). Therefore, the radioanalytical results of the three primordial radionuclides were used to quantify several radiological hazard indexes that are briefly introduced in the following subchapters.

Radium equivalent activity in soil
Radium equivalent (Ra eq ) is a radiological health risk parameter resulting from exposures to naturally occurring radioactive materials containing 232 Th, 238 U, and 40 K. It is used by regulatory bodies to set a single regulatory limit rather than setting individual radionuclide limits (Abed et al. 2022). The radium equivalent (Ra eq ) for soil samples was calculated using Eq. 3 (Tufail 2012): where A u , A Th , and A k are the activity concentrations (Bq kg -1 ) of radionuclides in soil samples.

External and internal hazard indexes
The H ex is a single index that measures the exposure to gamma radiation from 232 Th, 238 U, and 40 K radionuclides that are found naturally in soil. Using Eq. 4, the H ex was computed (Akpanowo et al. 2020). It is noteworthy that the soil is commonly used as a construction material to build earthen houses in the study area.
where A u , A Th , and A k are the activity concentrations (Bq kg -1 ) of radionuclides in soil samples. The internal hazard index (H in ) is a parameter used to assess the indoor radiological hazard from radionuclides in construction materials, including the exposure of radon and radon progeny originating from the radioactive decay of primordial radionuclides. The H in was calculated with Eq. 5: The external and internal hazard indices are acceptable (i.e., not exceeding the dose limit of 1 mSv y -1 ) when the calculated index is ≤ 1 (Purnama and Damayanti 2020).

Radioactivity level index for farmers
The radioactivity level index, noted by Iγ, has been proposed to estimate the magnitude of radiation risk from gamma emitting radionuclides in soil. It was estimated using Eq. 6 (Purnama and Damayanti 2020).
A U , A Th , and A K are the activity concentrations (Bq kg -1 ) in soils of 238 U, 232 Th, and 40 K, respectively. Radiation safety is considered acceptable if Iγ ≤1.

Absorbed dose rate
The absorbed dose rate by human beings (D) from exposure to external gamma radiation as a result of radioactive decay of 238 U, 232 Th, and 40 K contained in soils was estimated using the activity-to-dose conversion factors (from Bq kg -1 in soils to nGy h -1 absorbed dose) and the activity concentrations A U , A Th , and A K determined in soil samples by gamma spectrometry. The activity-to-dose conversion factors used were 0.462 ( 238 U), 0.621 ( 232 Th), and 0.0417 ( 40 K). The absorbed dose rate was calculated using Eq. 7 (UNSCEAR 2008): where D is the absorbed dose rate (nGy h -1 ) and A U , A Th , and A K are the activity concentrations (Bq kg -1 ) of 238 U, 232 Th, and 40 K in soil samples.

Annual effective dose for tobacco farmers from tobacco leaves
Through the root uptake of radionuclides from soils, tobacco plants can become an additional above ground reservoir of radionuclides. Therefore, we hypothesized that tobacco plantations (i.e., the plant biomass) might be a source of external radiation to farmers and workers of the plantations, and the relevant radiological indexes were calculated as follows.
Applying the ICRP activity-to-dose conversion factors (Sievert per gray (Sv/ Gy)), occupancy factors for indoor and outdoor (0.8 and 0.2, respectively), the time of a year (hours per year (h/y)), and the absorbed dose rate D (nGy h -1 ), the yearly effective dose was calculated (UNSCEAR 2008). The annual effective dose to tobacco farm workers outdoors (AED o ) and indoors (AED i ) were calculated using Eq. 8 and Eq. 9 (Taskin et al. 2009;Akpanowo et al. 2020): The total annual effective dose AED is the sum of AEDo and AEDi. It should not exceed an annual dose limit of 1 mSv y -1 to be considered acceptable for the general public.

Annual effective dose from tobacco inhalation
Tobacco can enter the human lungs through snuffing or smoking. A distinction was made as follows: the calculation of radiation dose from tobacco snuffing assumes that the gamma-emitting radionuclides present in tobacco reach the lungs. In contrast, in the case of smoking, only a fraction of the present radionuclides reach the lungs, and other fractions escape through the smoke main stream into the air or are retained by the cigarette filter.
The annual effective dose of tobacco leaf consumption by snuffing was calculated using Eq. 10.
(Eq. 10) where A is the activity concentration of individual radionuclides in tobacco leaves, M is the total mass of the snuffed tobacco in one year, and D cf is the activity-to-dose conversion coefficient (Sv Bq -1 ).
The annual effective dose inhaled from tobacco/cigarette smoke was calculated using Eq. 11.
It was estimated and assumed that about 75% of the radionuclides present in tobacco and tobacco products (cigars, cigarettes, etc.) pass into smoke. However, only 50% of that smoke reaches the lungs, while the other 50% is released into the air (side smoke) (UNSCEAR 2008). Therefore, the smokers and snuffers inhalation dose conversion coefficients for individual radionuclides used were 2.9 10 -6 Sv Bq -1 for 238 U, 4.5 10 -5 Sv Bq -1 Th 232 , and 2.1 10 -9 Sv Bq -1 for 40 K.
The annual effective dose from tobacco smoking and snuffing was calculated using local statistics available for cigarette consumption in Tanzania (GATS 2018). The calculations were based on the consumption of 8.5 cigarettes per day per smoker/snuffer (GATS 2018). The average weight of tobacco rolled by the smoker/snuffer in a piece of paper, banana leaf, maize leaf (traditionally known as "roll-your-own tobacco"), or proper cigarette paper is approximately 1.3 grams. Therefore, the weight of tobacco consumed annually in Tanzania was calculated considering the weight of tobacco in a cigarette multiplied by the number of cigarettes smoked per day (1.3 g × 8.5 = 11.05 g per day), which yields the annual mass of tobacco consumed of approximately 4 kg per person (11.05 g per day per person × 365 days per year = 4.033 kg y -1 ). Unfortunately, the information on partitioning the amount of tobacco consumed as smoked or snuffed per year was not available, and it was assumed that the amount of tobacco snuffed is equal to the amount smoked. This information was used in Eq. 10 and Eq. 11 to estimate the effective dose for tobacco consumers through inhalation.

Excess lifetime cancer risk (ELCR) for tobacco consumers
Adoption of the linear non-threshold (LNT) dose-response model proposed by the International Commission for Radiological Protection (ICRP) implies that low radiation doses from naturally occurring radioactive materials can potentially induce cancer in exposed individuals. Therefore, the risk of developing cancer from low doses over the person's lifetime or the excess lifetime cancer risk (ELCR) for smoking and snuffing was calculated using Eq. 12 and Eq. 13 (Kadhim and Ridha 2019).
(Eq. 11) where ELCR is the excess lifetime cancer risk, A ls is the average life expectancy (75 years) for snuffers and smokers, 0.05 Sv -1 is the mortality risk coefficient for inhalation, and E smoking and E snuffing are the annual effective doses from smoking and snuffing, respectively.
According to United Nations Commodity Trade Statistics Database (UN Comtrade) 2022, tobacco and tobacco products are consumed in the producing nations (Kenya, Tanzania, and Uganda), as well as other African nations (Burundi, Malawi, Rwanda, and South Africa), with average life spans of 65 years. Therefore, the ELCR from snuffed and smoked tobacco and tobacco products was calculated using an A ls value of 65 years, the average life expectancy across the three countries.

Statistical analyses
Following descriptive statistical evaluation using the Shapiro-Wilk normality test, the distribution of all data was determined to be normally distributed (p < 0.05). Then, further analysis was done using Tukey Honestly Significant difference (HSD) to show the differences between the means. Finally, the degree of the relationships between the radionuclide's concentration ( 238 U, 232 Th, and 40 K) and soil physical-chemical properties was determined using Pearson's correlation analysis (p < 0.05). The statistical analyses were performed using the STATISTICA (8 th Edition) software.

One-year experimental results
The activity concentrations of 238 U, 232 Th, and 40 K in the NPK fertilizers used in the one-year experimental plots are indicated in Table 2. These fertilizer brands are also the most commonly used in tobacco plantations in Kenya, Tanzania, and Uganda. The results show that the GLT fertilizer had the highest 232 Th concentration (687.7 ± 1.5 Bq kg -1 ) and the highest 238 U concentration (3,216.3 ± 21.7 Bq kg -1 ). The high radioactivity of the GLT fertilizer is associated with the radioactive nature of the local phosphate ore, which serves as the primary raw material in fertilizer production (Mwalongo et al. 2022). Potassium is one of the macronutrients present in the fertilizer. It mainly comprises the stable isotope 39 K, but it always contains the radioactive isotope 40 K in a fixed proportion, and 40 K makes up about 0.012% (120 mg kg -1 ) of the elemental potassium.
Compared to the other fertilizers, YC showed much lower amounts of 232 Th (26.3 ± 0.3 Bq kg -1 ) and 238 U (198.6 ± 6.1 Bq kg -1 ), indicating that the source material for YC phosphate fertilizer may have low radioactivity levels.
(Eq. 13) ELCR snuffing = A ls × E snuffing × 0.05Sv −1 Despite the low-level radioactive detection, the product (YC) is sufficiently suitable for use as fertilizers for tobacco. According to the much greater 40 K radionuclide concentration (1,566.7 ± 98.5 Bq kg -1 ) found in YB, the raw materials stable potassium ( 39 K) may include the radioactive 40 K (Anguani 2017). The crop leaf yield (kg ha -1 ) was affected by fertilizer application. Applying YB and YC increased leaf yield and 40 K-content significantly compared to leaves grown using GLT ( Table 2). The increase in tobacco leaf 40 K was due to the applied good quality NPK fertilizers.
Results from the one-year experimental plots are shown in Fig. 3. The concentrations of the 238 U, 232 Th, and 40 K activity in NPK-treated soils were significantly (p < 0.05) higher than the soil from the control plot where no fertilizers were applied (Fig. 3). The results also show that the soil treated with YB displayed significantly (p < 0.05) higher 40 K activity concentration than soil from the other plots. High 238 U levels in the soil are assumed to be caused by the high 238 U content of the phosphate rock used in fertilizer production. The soil treated with GLT fertilizer contained significantly (p < 0.05) higher 238 U activity concentration (30.57 ± 1.3 Bq kg -1 ) if compared with the other plots.
There were no significant differences in activity concentrations of 238 U and 232 Th in soils treated with YB and YC fertilizers (Fig. 3). Low 238 U and 232 Th in YB and YC fertilizer is attributed to low 238 U and 232 Th in the source rock or the radionuclides could have been washed out by rainfall. Furthermore, the Tukey's Honestly Significant Difference (THSD) post hoc test revealed that the activity concentrations of 238 U, 232 Th, and 40 K after fertilizer application did not vary significantly with soil depth, i.e., the tillage of the soil distributed the fertilizers in the 30 cm topsoil layer.
In Fig. 3, the letters indicate the Tukey significant difference among the average concentrations of 232 Th, 238 U, and 40 K across the column. In the Tukey's multiple pairwise honestly significant (THSD) statistical test, the different letter(s) on each similar bar chart are significantly different at p ≤ 0.0 5 confidence level.  Table 2, the letters indicate the Tukey significant difference among the average concentrations of 232 Th, 238 U, and 40 K across the column. The Tukey's multiple pairwise honestly significant (THSD) statistical test, the column with different letter(s) are significantly different at p ≤ 0.0 5 confidence level. Meaning of the letters: "a" is very highly significant; "b" is very significant; "c" is significant; "d" is non-significant The activity concentrations of 232 Th, 238 U, and 40 K in tobacco leaves from the one-year experimental plots are shown in Fig. 4. The activity concentrations of 232 Th, 238 U, and 40 K in tobacco leaves from plots treated with NPK fertilizers were in the ranges of 5.97 ± 0.36 Bq kg -1 (GLT) to 74.45 ± 2.8 Bq kg -1 (YC) for 232 Th, 24.56 ± 0.78 Bq kg -1 (YB) to 28.58 ± 1.60 Bq kg -1 (GLT) for 238 U, and 1438.79 ± 12.1 Bq kg -1 (GLT) to 1539.38 ± 52.6 Bq kg -1 (YB) for 40 K. The results showed that the leaves of tobacco plants grown with NPK fertilizers displayed significantly higher 238 U concentrations than those of the control plants. The tobacco leaves showed a similar concentration of 232 Th for YB and GLT but a statistically higher 232 Th concentration for the YC fertilizer.
The 238 U uptake was significantly highest for tobacco leaves applied with GLT, followed by YC and YB. Conversely, the unfertilized tobacco leaves (control) had the significantly lowest uranium content (see Fig. 4). It is also noteworthy that the fertilized tobacco leaves showed considerably higher radionuclide concentrations than tobacco leaves analyzed elsewhere (Söğüt et al. 2014;Landsberger et al. 2015).
The activity concentrations of 40 K in tobacco leaves were significantly higher than those of 238 U and 232 Th (Fig. 4) in plots treated with YB and YC fertilizers. The higher concentration of 40 K is likely due to the hyperabsorption of potassium by tobacco plants in comparison with other macronutrients (Çalişkan and Çalişkan 2018; Eke and Ishfaq 2021). The activity concentrations of 232 Th were significantly higher in tobacco leaves from plots treated with YC fertilizer compared with GLT, YB, and the control group. Except for the plot treated with YB fertilizer, the activity concentrations of 238 U and 232 Th in soils were still lower than the worldwide average concentrations of 238 U (35 Bq kg −1 ) and 232 Th (30 Bq kg −1 ) in soils reported by the United Nations Scientific Committee on Atomic Radiations estimates (UNSCEAR 2008). These experimental results show that when NPK fertilizers with high 238 U, 232 Th, and 40 K concentrations are applied, a transfer of these radionuclides from the soil to the tobacco plant is taking place.
The physical and chemical properties of soil from the oneyear experimental plots were determined, and the relationship between 232 Th, 238 U, and 40 K activity concentrations and soil physical-chemical properties was investigated. The results of a simple correlation matrix between the radionuclide concentrations and the physical-chemical properties is provided in Table 3. The correlation matrix revealed a moderately positive relationship between 232 Th and the soil sand fraction, and a moderately negative relationship with electrical conductivity, silt and clay fractions in soil, and an extremely weak relationship with soil pH. Thus, a strong positive correlation of 238 U with sandy soils could also reflect the ability of tobacco plant preference in sandy soils (Lisuma et al. 2020(Lisuma et al. , 2021. The 238 U displayed a strong positive correlation with sandy soil and a strong negative correlation with EC, silt and clay in the soil, and a moderate positive correlation with soil pH and organic carbon content. Furthermore, the 238 U concentration displayed a strong positive correlation with 232 Th and is moderately correlated with the 40 K concentration. Potassium ( 40 K) displayed a strong negative correlation with the soil pH, a moderately negative relationship with electrical conductivity, silt and clay fractions in soil, and an extremely weak relationship with organic carbon.
It has been reported that the behavior of uranium and other radionuclides in tropical soils depends upon radioelement solubility, soil pH, redox potential, cation exchange capacity (CEC), organic matter content, and soil texture (sandy, silted, and loamy) (Ribeiro et al. 2018;De Souza Braz et al. 2021). Similarly, according to Hu et al. (2020), soil physicochemical characteristics such as pH for sunflowers and peas showed the highest uranium uptake at pH 3-5. Another factor that affects uranium uptake by plants is the presence of carbonates and phosphates in the soil that have the propensity to interact with uranium to create complexes, which affect the absorption of uranium by plants (Sokolik et al. 2020;Tagami and Uchida 2020).

Ten-year NPK fertilizer application experiment
The results showed that the concentration of 232 Th, 238 U, and 40 K in agricultural soils used in this experiment was significantly higher (p < 0.05) after ten years of annual application of phosphate fertilizers than the control samples (no application of fertilizer) (Figs. 5 and 6).
The higher 238 U concentration in Manyoni was related to the prolonged phosphate fertilizer application, but the naturally occurring uranium in soils in this area is also slightly above average (Kasoga et al. 2015). Indeed, the Manyoni region hosts some surficial uranium occurrences, which may have contributed to the higher radioactivity in local agriculture soils if compared to other farms (IAEA/NEA 2020). The activity concentrations of 238 U, 232 Th, and 40 K in the agricultural soils are shown in Figs. 5 and 6.
The results indicate that 232 Th was not detected in tobacco leaves from the Biharamulo and Bukombe farms meaning that the value was below the method's detection limit of 1.56 ± 0.12 Bq kg -1 . Tobacco leaves from Kahama (33.73 ± 1.60 Bq kg -1 ) showed the highest 232 Th concentration, followed by those from the Urambo farm (26.02 ± 1.50 Bq kg -1 ), Kakonko (15.73 ± 1.42 Bq kg -1 ), Manyoni (12.83 ± 1.03 Bq kg -1 ), and Namtumbo (12.43 ± 0.76 Bq kg -1 ). The range of 232 Th concentrations in tobacco leaves reported in this study agrees well with those reported by Mkhaiber et al. (2020), who assessed the presence of 232 Th in tobacco leaves from the UAE, France, Iraq, Armenia, Türkiye, and South Korea and reported 232 Th concentrations ranging from 4.8 Bq kg -1 to 36.0 Bq kg -1 .
The results show that the concentration of 232 Th decreased from Kahama down to Urambo, Kakonko, Manyoni, and Namtumbo (located in the southern part of Tanzania). The 232 Th decreasing trend may be attributed to the variation of the physical-chemical properties of the soils. Thus, an increase in acidity from Kahama to Namtumbo influenced the increase of 238 Th. A similar observation was reported by Gupta et al. (2020), whose results showed an increase of Th in more acidic soils and U decreasing in acidic soils. 1 3 The concentration of 40 K (Fig. 8) was significantly higher in tobacco leaves from Biharamulo (1,545.30 ± 58.87 Bq kg -1 ) and Kahama (1,538.14 ± 32.3 Bq kg -1 ) followed by Bukombe (1,497.83 ± 78.89 Bq kg -1 ), Kakonko (1,434.3 ± 78.7 Bq kg -1 ), Urambo (1,255.1 ± 60.9 Bq kg -1 ), Manyoni (1,081.8 ± 46.6 Bq kg -1 ), and Namtumbo (823.9 ± 75.9 Bq kg -1 ). However, the 40 K concentration of Namtumbo was statistically similar to the control sample.
The physical-chemical soil factors influence the concentration of 238 U, 232 Th, and 40 K in the soil. Studies have shown that the geology and soil structure influence environmental concentration of radionuclides of natural origin (UNSCEAR 2008). The concentration of 238 U, 232 Th, and 40 K can also be influenced by soil weathering, sediment accretion, sorption, leaching, and circulation of groundwater and surface water (Gupta and Walther 2019;Ratnikov et al. 2020). All these factors certainly played a role in the varying radionuclide concentrations among regions, as found in this investigation. In the ten (10) year's experiment, it was, however, impossible to determine the exact role of each factor. Nevertheless, it was found that there is a strong and positive correlation between 238 U and 232 Th, and both displayed a strong negative correlation with organic matter in the soil. Furthermore, the sandy soil was positively correlated with 238 U and 232 Th and negatively correlated with organic matter and silt, as indicated in Table 4.

Investigation in selected tobacco farms in Kenya, Tanzania, and Uganda
The activity concentration of 232 Th, 238 U, and 40 K was determined in the agricultural soils of farms in Kenya, Tanzania, and Uganda for the soil layer with 0-30 cm depth. The concentrations of 232 Th, 238 U, and 40 K in soils were significantly higher (p < 0.05) than the control sample from the non-fertilized farm in Tabora, Tanzania (Figs. 9 and 10). Migori (Kenya) had a significantly higher 238 U concentration, followed by Kanungu (Uganda), which showed the lowest value that was still significantly higher than the control group. A similar result trend was observed for 232 Th (Fig. 9). However, for 40 K, Kanungu (Uganda) had a significant concentration, followed by Migori (Kenya), and Urambo (Tanzania) had the lowest 40 K concentration (Fig. 10), which was similar to the control group. The variations in natural radioactivity in soil depend on the unique natural radionuclides present at the sampling sites, the parent rock that formed the soil, and other forming variables (UNSCEAR 2008).
Tobacco leaves from Kanungu showed the highest 40 K concentration (1247.5 ± 24.0 Bq kg -1 ), followed by In these three countries, the radionuclide concentrations in tobacco leaves were globally higher than in the 1-year and 10-year experiments for 238 U and 232 Th. However, the 40 K concentrations were even lower than that of the control sample.
The Pearson's correction matrix on the association between 232 Th, 238 U, and 40 K concentrations and the physical-chemical parameters for agricultural soils of Kenya, Tanzania, and Uganda are presented in Table 5. The results showed a negative association of sand and silt with that of OM, indicating that the OM content was higher in clayey soils. The negative association may be because sand and silt lose OM more rapidly than clayey soils through natural and anthropogenic processes (Arunrat et al. 2020). Another possible reason is that indigenous tobacco farmers uproot stalks from the field after harvest and burn them, thus reducing organic carbon from the decay of plant materials in soils (Lisuma et al. 2022b). This result is similar to that Lisuma et al. (2022a) reported earlier, who found that Tabora and Urambo soils are mainly composed of sand and silt and contain the lowest OM content.
The results further indicate a strong positive association between 238 U and OM. The association between 238 U and OM results from soil OM reducing the water-soluble and mobile hexavalent uranium U(VI) to the insoluble tetravalent U(IV), and thus, over time, the U concentrations in soils increase (Sokolik et al. 2020). There was a positive correlation between the 40 K concentration and the soil pH and a negative correlation between 40 K and the clay soil fraction.

Radium equivalent activity of soils
The Ra eq of agricultural soils from all study areas ranged from 68.17 ± 3.5 Bq kg -1 (Urambo) to 329.2 ± 12.8 Bq kg -1 (Manyoni) and is shown in Table 6. These values are below the Ra eq value of 370 Bq kg -1 , corresponding to delivering an annual dose of 1 mSv y -1 to public members (UNSCEAR 2008). The low Ra eq implies that these soils can be used as a construction material, as is common in the investigated regions.

Absorbed dose rate due to radioactivity in soils
The obtained results show that the absorbed dose rate from agricultural soils ranged from 32.38 nGy h −1 (Urambo) to 282.93 ± 14.67 (Bukombe) (see Table 6). The absorbed dose rate for the Urambo (32. 38 nGy h −1 ), the Yaramila compound experimental site (33.14 nGy h −1 ), Kakonko (40.5 nGy h −1 ), Kanungu (41.44 nGy h −1 ), and Migori (54.37 nGy h −1 ) were lower than the global average value of 60 nGy h −1 (UNSCEAR 2008). Therefore, farmers and public members using these lands are exposed to low doses of gamma radiation from the agricultural soils. However, the absorbed dose rate at the Tanzania

The annual effective dose rate from radioactivity in soils
The calculated outdoor and indoor annual effective dose rates from soils in the tobacco plantations, as summarized in Table 6, ranged from 37.0 nSvh -1 (GLT experimental site) to 190.0 nSvh -1 (Manyoni) and 150 nSvh -1 (GLT, YC, Urambo, Kakonko) to 750 nSvh -1 (Manyoni), respectively. The annual effective dose rates were below the worldwide mean annual effective dose, which is about 500 μSv y −1 . Moreover, the total annual effective doses have been found to vary depending on the geological parameters of the soil matrixes where tobacco plants are grown and are further influenced by anthropogenic activities such as applying mineral phosphate fertilizers (IAEA 2013;UNSCEAR 2008).

The external and internal hazard indexes
The results in Table 6 show that the H ex and H in of tobacco agricultural soil ranged from 0.18 (YC, Urambo) to 0.89 (Manyoni) and 0.23 (Urambo) to 1.08 (Manyoni), respectively. For the radiation hazard to be tolerable, it should be lower than 1, which is (mostly) the case here. An index result greater than 1 corresponds to an annual dose of more than 1 mSv y -1 (Imani et al. 2021). The results imply that if the soil is used as a building material, this practice can exceed the public annual effective dose constraint of 1 mSvy -1 (UNSCEAR 2008).

Tobacco leaf radium equivalent activity
Tobacco plants were assumed to constitute an above ground reservoir of 232 Th, 238 U, and 40 K and thus a potential source of radiation exposure to human beings and surface soils. Accordingly, the values of Ra eq with the lower value of 90.20 ± 4.6 Bq kg -1 in Namtumbo and the highest value of 174.50 Bq kg -1 at the Kahama sampling location were measured (Table 7). Compared to the worldwide average of 60 Bq kg -1 of Ra eq , all investigated tobacco plant leaves exceeded the worldwide average Ra eq value for soils. Nevertheless, the equivalent radium activity ranged from 90.20 Bq kg -1 to 174.5 Bq kg -1 , which is significantly lower than the UNSCEAR recommended combined activity concentration of 370 Bq kg -1 (UNSCEAR 2008; Mostafa et al.

2020)
, that again corresponds to an annual dose of 1 mSv y -1 for the general public. Therefore, the Ra eq activity indicates that the tobacco crops do not pose a significant radiological risk to farm workers and relatives living nearby.

Annual effective dose due to tobacco inhalation
The results shown in Table 7 indicate that the annual effective dose due to tobacco snuffing ranged from 0.01 to 6.53 mSv y -1 , varying with the producing region. The tobacco leaves from the one-year experimental plots may lead to effective doses higher than the recommended safety limit value of 1 mSv y -1 if snuffed and may pose a severe radiological health hazard.

3
The annual effective dose equivalent for tobacco smoking ranged from 0.1 mSv y -1 to 2.45 mSv y -1 , varying again with the tobacco producing regions. Tobacco smokers may receive an annual dose that is 50-194% higher than the recommended limit of 1.26 mSv y -1 , the average annual global exposure to natural radiation sources from inhalation (UNSCEAR 2008;Zlobina et al. 2022). Therefore, snuffing and smoking tobacco can produce relatively high annual effective doses.

Excess lifetime cancer risk
The excess lifetime cancer risk (ELCR) values for tobacco snuffing ranged from 5.0 × 10 -5 to 24.48 × 10 -3 for snuffing and from 0.02 × 10 -3 to 9.18 × 10 -3 for smoking, respectively. Therefore, as shown in Table 7, tobacco snuffing and smoking increase the likelihood of cancer significantly if compared to the worldwide average exposure of non-smokers/snuffers.

Conclusions
The activity concentrations of 232 Th, 238 U, and 40 K were determined in the three NPK fertilizers used in growing tobacco crops in Kenya, Tanzania, and Uganda. The results showed that YB and GLT fertilizers had higher concentrations of 232 Th, 238 U, and 40 K in soils than the worldwide average of 238 U (35 Bq kg −1 ) and 232 Th (30 Bq kg −1 ) in soils reported by UNSCEAR. In addition, a one-year controlled experiment revealed that NPK fertilizer application caused soil activity concentrations of 232 Th, 238 U, and 40 K and subsequently increased tobacco leaves uptake compared to plots where no mineral fertilizers were applied (control).
The application of phosphate fertilizers for a longerterm (10 years experiment in this study) also increased concentrations of 232 Th, 238 U, and 40 K in soils and harvested tobacco leaves compared to the unfertilized tobacco farms. Thus, selecting NPK fertilizer with a lower content of naturally occurring radionuclides could reduce agricultural soil contamination and reduce radionuclide soil-to-plant transfer and, therefore, reduce the radiological health impact through inhalation and ingestion to humans.
The results showed a clear relationship between the 232 Th, 238 U, and 40 K concentrations and physical-chemical properties which varied according to the soil's physical and chemical properties. However, the results also indicate that the concentration of the radionuclides was not correlated with the soil's physical-chemical properties.
Radiation hazard parameters such as the radium equivalent activity, gamma-absorbed dose, annual effective dose, external and internal hazard indexes, and gamma index to both soils and tobacco leaves to assess the radiological risk parameters for the public, in general, were determined in this work. The results show that all radiological health risk parameters resulted in values lower than the recommended safety limits. Hence, exposure to agricultural soil and tobacco crops fertilized with phosphate fertilizers poses radiological hazards to farm workers and members of the public that are within tolerable limits. Even using these soils as construction materials for rural houses (a common practice in the study area) does not pose a radiological threat to the population. Therefore, the current work recommends further studies to expand the knowledge on soil-to-plant radionuclide transfer and the role of fertilizers in such transfers in cereal crops since cereal crops are grown in rotation with tobacco plantations in the investigated regions. Even if the measured radioactivity levels are within tolerable limits, it is recommended that countries should use fertilizers with lower radioactivity levels for the conservation of soil quality and to reduce radionuclide concentrations in tobacco.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.