The study, published in the journal Environmental Monitoring and Assessment, was carried out by Isaiah Eze Igwe and Usman Ibrahim Machika of the Federal University Dutsin-Ma, together with Chibueze Anslem Mgbechidimma of the University of Ibadan. Their target was the dumpsite of the NAK Steel company in Katsina State, an environment typical of many industrial waste grounds across the region, where slag, scrap, and contaminated soils create a patchwork of radiation levels that conventional single-point surveys can easily miss. The researchers assembled a dataset of 36 soil sampling locations across the site and used it to train and compare four supervised learning algorithms: support vector machine, linear regression, random forest, and decision tree.

The choice of algorithms was deliberate. Linear regression represents the simplest statistical baseline, fitting a straight-line relationship between input variables and the measured dose rate. Support vector machines, first introduced in the 1990s, are more flexible, capable of capturing nonlinear patterns by mapping data into higher-dimensional spaces. Decision trees split the data through a cascade of binary decisions, producing piecewise predictions that can adapt to sharp local changes. Random forests, an ensemble method pioneered by Leo Breiman in 2001, combine hundreds of decision trees, each trained on a random subset of the data, and average their outputs. This averaging tends to suppress the overfitting that plagues individual trees while preserving the ability to model complex, nonlinear spatial relationships.

All four models were trained and tested under identical preprocessing and validation conditions, ensuring a fair comparison. The results were decisive. Random forest outperformed every competitor across all performance metrics, achieving a root mean square error of just 0.0131 and a mean absolute error of 0.0101, the lowest prediction errors of the group. Its coefficient of determination, a measure of how much of the variability in the data the model explains, reached 0.902, meaning the model accounted for roughly ninety percent of the spatial variation in absorbed gamma dose rate. When the task was reframed as classifying locations into hotspot and non-hotspot categories, the random forest again led the field, with a classification accuracy of 0.89, a Cohen’s kappa of 0.77, and an area under the receiver operating characteristic curve of 0.93, all indicating strong and reliable discrimination.

The behavior of the other algorithms was just as instructive, though for different reasons. Linear regression and support vector machines both tended to smooth out spatial variability, producing predictions that underestimated localized anomalies. In a radiation context, that smoothing is not a harmless simplification; it can hide exactly the concentrated pockets of elevated dose that matter most for protecting workers and nearby residents. The decision tree model showed the opposite failure mode, generating abrupt spatial patterns suggestive of overfitting, where the model memorized the quirks of the training samples rather than learning the underlying structure of the radiation field. The contrast highlights a central lesson of modern machine learning applied to environmental data: the best model is rarely the most flexible or the simplest, but the one whose inductive bias matches the phenomenon being mapped.

With the best-performing model in hand, the researchers converted their predicted dose rates into two internationally recognized radiological risk indicators. The first was the annual gonadal dose, an organ-specific measure of radiation exposure to the reproductive organs, which is of particular concern because reproductive tissue is among the most radiosensitive in the human body. Predicted annual gonadal dose values across the site ranged from 0.03 to 0.27 millisieverts per year. The second metric was the excess lifetime cancer risk, a probabilistic estimate of the additional cancer burden that long-term exposure to the site’s radiation field would impose on a hypothetical resident. The ELCR estimates mirrored the spatial variability of the annual effective dose, rising and falling in step with the predicted dose rates across the dumpsite.

The spatial pattern of risk was far from uniform. The highest values of both dose and cancer risk clustered around a single location designated S11, identified as the principal hotspot of the site. This concentration is exactly the kind of feature that sparse or averaged survey campaigns can overlook, and it underscores why spatially resolved prediction matters. At a working dumpsite, where scavengers, workers, and nearby communities may spend hours in specific zones rather than wandering uniformly across the whole area, knowing where the hotspots lie transforms a vague site-wide average into actionable information: which zones to fence, which to remediate, and which to monitor most frequently.

The study is candid about its limitations. A dataset of 36 sampling points is small by machine learning standards, and the authors acknowledge that this constrains the generalization of their models beyond the specific site studied. Small training sets carry a real risk that a model’s apparent accuracy partly reflects the particular configuration of the validation split rather than genuine predictive power. Nevertheless, the framework’s value lies less in the specific numbers it produced for NAK Steel and more in the demonstration that a rigorous, low-cost predictive pipeline, from soil sampling through algorithm comparison to dose and risk conversion, can function in data-sparse regions where expensive radiation monitoring infrastructure is limited. The authors followed internationally accepted conversion models in moving from predicted dose rates to organ doses and lifetime risk, anchoring the machine learning output in established radiological protection practice.

The broader context makes the work timely. Industrial dumpsites across Nigeria and much of the developing world are increasingly recognized as potential hotspots of environmental radiation exposure, both because industrial byproducts such as slag can concentrate naturally occurring radioactive materials and because unregulated waste handling allows contamination to spread. Previous studies at Nigerian dumpsites and mining areas have documented elevated gamma dose rates and raised concerns for waste pickers, whose informal and unprotected work places them in direct and prolonged contact with contaminated soil. Yet predictive assessments of dose distribution in these settings have remained scarce, leaving regulators with little more than point measurements to guide decisions.

What this study offers is a practical template. By showing that random forest can reliably predict absorbed gamma dose rate in a heterogeneous industrial environment, and by chaining that prediction into organ dose and cancer risk estimates, the researchers have sketched a workflow that other teams can replicate at other sites with modest sampling budgets. The approach also aligns with a growing international trend of applying machine learning to environmental radioactivity, from mapping geogenic radon potential to modeling soil-plant transfer of radiocaesium. For data-sparse regions, where the choice has often been between sparse surveys and no assessment at all, predictive modeling offers a middle path: a way to extract maximum insight from limited measurements, direct future sampling to where it matters most, and give communities and regulators a clearer picture of the invisible radiation landscape they live and work within.

Subject of Research: Machine learning prediction of absorbed gamma dose rate and radiological risk assessment at an industrial steel dumpsite in Nigeria

Article Title: Machine learning prediction of absorbed gamma dose rate and radiological hazard assessment at a Nigerian steel dumpsite

Article References: Igwe, I. E., Machika, U. I., & Mgbechidimma, C. A. (2026). Machine learning prediction of absorbed gamma dose rate and radiological hazard assessment at a Nigerian steel dumpsite. Environmental Monitoring and Assessment, 198(10), Article 1046. https://doi.org/10.1007/s10661-026-15865-2

Image Credits: AI Generated

DOI: 10.1007/s10661-026-15865-2

Keywords: machine learning, random forest, gamma dose rate, radiological risk assessment, industrial dumpsite, Nigeria, annual gonadal dose, excess lifetime cancer risk, environmental radiation monitoring, soil radioactivity, hotspot mapping, Katsina State

The research team, led by Lawal Zubairu of the Sino-French Institute of Nuclear Engineering and Technology at Sun Yat-sen University, together with colleagues at the Nigeria Atomic Energy Commission’s training institute, collected thirty soil samples from across the Rafin Gora mining area. Back at the laboratory of the National Animal Production Research Institute in Zaria, the samples were digested and analyzed by flame atomic absorption spectrophotometry, a workhorse technique that atomizes the sample in a flame and measures how much light each metal absorbs, allowing precise quantification of cadmium, cobalt, chromium, copper, manganese, nickel, lead, arsenic, and zinc. The team then ran the numbers through a battery of established pollution indices and the United States Environmental Protection Agency’s human health risk models.

The pollution indices tell a story of contamination that is real but patchy. The contamination factor, which compares the concentration of each metal in a sample to its background level, revealed that cadmium, arsenic, and zinc showed localized enrichment—meaning that at certain sampling points, human activity had clearly elevated these metals above natural levels. The geoaccumulation index, a logarithmic scale introduced by Müller in 1969 to grade pollution intensity in sediments, classified more than ninety percent of the samples as unpolluted. Yet the pollution load index, which integrates the contamination factors of all measured metals into a single site-level value, exceeded one at most locations, a threshold that signals overall anthropogenic influence rather than purely geological origin.

Ecologically, the news is relatively reassuring. The ecological risk index, originally developed by Hakanson in 1980 for aquatic pollution control, weighs each metal’s toxicity and abundance to estimate the potential harm to ecosystems. Across all sampling locations at Rafin Gora, this index indicated low ecological risk. In other words, the soil ecosystem itself is not yet in crisis. But ecological risk and human health risk are not the same thing, and it is on the human side of the ledger that the study’s most consequential findings emerge.

Using the USEPA’s risk assessment framework, the researchers calculated chronic daily intake values for each metal through three exposure pathways: ingestion of soil particles, inhalation of dust, and dermal contact with soil. The result was unambiguous. Ingestion dominated as the exposure route, and children absorbed doses five to ten times higher than adults. This is not surprising to exposure scientists: children play on the ground, put dusty hands and objects in their mouths, and ingest far more soil per kilogram of body weight than adults do. In a mining village where tailings and crushed ore are scattered across living spaces, that behavioral difference translates directly into a chemical dose difference.

When the researchers summed the hazard quotients—the ratio of each metal’s estimated intake to its reference dose, the level below which no adverse effect is expected—into a hazard index, the values came out at 0.026 for adults and 0.139 for children. Both sit below unity, the conventional line separating acceptable from unacceptable non-carcinogenic risk. Within those totals, however, chromium and arsenic contributed the largest shares of both the hazard quotient and the intake, marking them as the metals that matter most at this site. Cadmium, lead, and the others, despite their fearsome reputations, played comparatively minor roles in the overall risk picture at Rafin Gora.

The carcinogenic risk calculation is where the study acquires its edge. Arsenic and chromium are both recognized human carcinogens, and the USEPA models assign them slope factors that convert chronic intake into a lifetime probability of cancer. The total carcinogenic risk came to 3.66 × 10⁻⁶ for adults and 6.79 × 10⁻⁶ for children. To put those numbers in context: the widely used acceptable upper limit for regulatory purposes is 1 × 10⁻⁴, or one additional cancer case per ten thousand people, and both values fall comfortably below it. But the threshold of negligible risk is conventionally set at 1 × 10⁻⁶, one in a million, and both values—particularly the children’s figure, nearly seven times that level—exceed it. The authors conclude that locally enriched arsenic and chromium drive a non-negligible cancer risk, especially for children.

The Rafin Gora findings resonate with a broader and often tragic pattern across Nigeria’s artisanal gold mining belt. Artisanal and small-scale gold mining is a lifeline for millions of people in West Africa, but it frequently proceeds without environmental safeguards, leaving crushed ore, tailings, and contaminated dust in intimate contact with communities. Previous Nigerian studies have documented heavy metal contamination around artisanal sites in Zamfara, Niger, Osun, and Nasarawa states, and the 2010 Zamfara lead poisoning epidemic—which killed hundreds of children and was traced to lead-rich ores processed inside village compounds—remains one of the starkest demonstrations of what unregulated ore processing can do. The Rafin Gora study adds a carefully quantified data point to this regional picture, showing that even where acute poisoning is absent, chronic low-level exposure can quietly accumulate statistical risk.

Methodologically, the study also illustrates both the power and the limits of standard risk assessment. The USEPA models rely on default exposure assumptions—soil ingestion rates, body weights, exposure frequencies—that are drawn largely from Western datasets and may not perfectly match the realities of rural Nigerian life. Moreover, total metal concentration is not the same as bioavailable metal: only the fraction of a metal that dissolves in the gut can actually be absorbed, and that fraction varies with the metal’s chemical form, or speciation. The authors explicitly recommend that future work incorporate bioaccessibility testing and speciation analysis, which would refine the risk estimates and could either temper or sharpen the current conclusions. They also call for practical mitigation at the site itself, including dust suppression to cut the dominant ingestion and inhalation pathways, and soil remediation to remove or immobilize the enriched arsenic and chromium.

For the people of Rafin Gora, the message of this research is neither panic nor complacency. The soils of the mining area are, by the standard indices, mostly unpolluted, the ecological risk is low, and the non-cancer hazard indices sit safely below regulatory thresholds. Yet the same measurements reveal pockets of arsenic and chromium enrichment that push lifetime cancer risk above the level scientists consider negligible, with children bearing the heaviest burden. In communities where gold mining is both livelihood and hazard, that combination argues for targeted, inexpensive interventions—wetting down dusty paths, keeping children’s play areas clear of tailings, remediating the worst hotspots—rather than wholesale abandonment of the mines. It is a reminder that in environmental health, the dose and the exposed population matter as much as the poison, and that careful measurement is the first step toward protecting the most vulnerable.

Subject of Research: Heavy metal contamination and human health risk assessment in soils of an artisanal gold mining area in Katsina State, Nigeria

Article Title: Health risk assessment of heavy metals exposure in soil at Rafin Gora artisanal gold mining area, Nigeria

Article References: Zubairu, L., Mingliang, K., Xueqing, Y., & Raza, K. (2026). Health risk assessment of heavy metals exposure in soil at Rafin Gora artisanal gold mining area, Nigeria. Environmental Science and Pollution Research, 33(30), 15865-15882. https://doi.org/10.1007/s11356-026-38155-3

Image Credits: AI Generated

DOI: 10.1007/s11356-026-38155-3

Keywords: heavy metals, artisanal gold mining, soil pollution, health risk assessment, arsenic, chromium, Nigeria, children's health, carcinogenic risk, USEPA models, environmental pollution, Katsina State