The research team, led by Kayode Olabode of Adam Mickiewicz University together with Karolina Lewińska and Izabela Komorowicz, focused on the Radość allotment-garden complex, where residents grow vegetables and spend leisure time on land enriched by historical ore extraction, roasting, and disposal of arsenic-bearing residues. Thirty-six surface soil samples were collected from the top 20 centimeters and analyzed for total arsenic using microwave-assisted digestion followed by inductively coupled plasma mass spectrometry. The results ranged from 433 to 1,148 milligrams of arsenic per kilogram of soil, with a mean of 698 milligrams per kilogram — a figure that dwarfs the Polish soil-quality threshold of 20 milligrams per kilogram for agricultural and recreational land. Previous studies in the region have reported even higher values, including up to 8,800 milligrams per kilogram in mine spoils and 17,200 milligrams per kilogram in tailings.

What makes this study technically distinctive is its treatment of bioaccessibility — the fraction of arsenic that actually dissolves in the acidic environment of the human stomach and is therefore potentially available for absorption into the bloodstream. The researchers used the Simplified Bioaccessibility Extraction Test, a validated in vitro method that simulates gastric juice with a glycine solution at pH 1.5, incubated at body temperature for one hour. The measured in vitro bioaccessible arsenic fraction ranged from 4.5 to 23.5 percent, with a mean of about 11 percent. This low bioaccessibility suggests that most of the arsenic in these soils is locked in relatively insoluble mineral phases, likely reflecting the weathering of arsenopyrite and the retention of released arsenic by iron and aluminum oxyhydroxides or its transformation into stable secondary arsenate minerals. In conventional risk assessments that assume 100 percent bioavailability, this distinction is ignored entirely — and, as the study shows, it matters enormously.

To translate soil chemistry into human health risk, the team evaluated three exposure pathways: incidental ingestion of soil, dermal contact with soil particles, and inhalation of resuspended dust. Rather than relying on a single deterministic calculation with fixed input values, the researchers ran a one-dimensional Monte Carlo simulation with 200,000 iterations for each receptor group, sampling arsenic concentrations and exposure factors from probability distributions. A key methodological innovation was the use of a joint paired smoothed empirical bootstrap, which preserved the measured statistical relationship between total arsenic and bioaccessibility during resampling, ensuring that simulated soil profiles reflected real-world pairing rather than artificial combinations of independent values.

The probabilistic results paint a concerning picture, particularly for children. Under the total-arsenic multi-pathway model, the median hazard index — a summary measure of non-cancer risk — was 5.50 for children, well above the screening threshold of 1, while the adult median was 0.62, though the upper tail of the adult distribution also crossed the threshold. The 95th percentile hazard indices reached 15.18 for children and 1.89 for adults. Cancer risk estimates were even more striking: the median total cancer risk was 1.78 × 10⁻⁴ for children and 1.06 × 10⁻⁴ for adults, both above the conventional 10⁻⁴ threshold at which risk is considered unacceptable. Critically, 80 percent of simulated child exposure profiles and 54 percent of adult profiles produced cancer risk estimates at or above that threshold. Even the 5th percentile values — representing the least-exposed end of the simulated population — exceeded the 10⁻⁶ safety floor for both groups.

The pathway analysis revealed a clear hierarchy of exposure. Ingestion dominated, accounting for 83 percent of children’s average daily dose and 71 percent of adults’ under the probabilistic model, with dermal contact contributing most of the remainder. Inhalation of resuspended particles was negligible, contributing essentially nothing to the modeled risk. Children’s higher doses stem from their frequent hand-to-mouth behavior and their greater soil intake relative to body mass — per kilogram of body weight, children eat, drink, and breathe more than adults, making them systematically more vulnerable to environmental toxicants. The deterministic point estimates generally fell between the probabilistic medians and 95th percentiles, a structural consequence of the right-skewed distributions produced when variable arsenic concentrations are combined with lognormal exposure factors.

The bioaccessibility refinement produced one of the study’s most consequential findings. When the measured in vitro bioaccessible fraction was applied only to the ingestion pathway — using identical soil and exposure-factor draws within each paired simulation so that only the concentration basis differed — the median ingestion-related hazard quotient and cancer risk dropped by 89.3 percent. For children, the proportion of simulations with ingestion hazard quotients above 1 fell from 93.3 percent to 8.7 percent, and the share with cancer risk at or above 10⁻⁴ collapsed from 51.9 percent to 0.6 percent. This means that assuming total arsenic is fully available for absorption overestimates ingestion risk by roughly ninefold. Yet the researchers caution that the refined estimates do not signal safety: the median child cancer risk remained above 10⁻⁵, and more than half of child simulations still fell in the 10⁻⁵ to 10⁻⁴ range, supporting continued exposure-control measures.

Sensitivity analysis using partial rank correlation coefficients identified the principal drivers of modeled risk. Total soil arsenic concentration and the soil ingestion rate emerged as the strongest positive predictors, with correlation coefficients approaching 0.9 for both children and adults, while exposed skin area and exposure duration also contributed meaningfully. Body weight was inversely associated with risk — lighter individuals receive higher doses per unit body mass — and inhalation rate was nearly irrelevant. This pattern carries a direct practical implication: because risk scales so strongly with soil concentration, mitigation strategies should prioritize reducing the amount of arsenic in the soil itself rather than focusing on behavioral factors alone.

The study also derived preliminary remedial goals — soil arsenic concentrations that would keep lifetime cancer risk below specified targets. At a target risk of 10⁻⁵, the conservative child 5th-percentile remedial goal was 13.0 milligrams per kilogram under the total-arsenic model, rising to 34.8 milligrams per kilogram after bioaccessibility adjustment of the ingestion pathway. Poland’s regulatory value of 20 milligrams per kilogram sits between these two estimates: more protective than the bioaccessibility-adjusted goal would require, but less conservative than the unadjusted child-based calculation. Either way, both remedial goals remain far below the measured soil concentrations of 433 to 1,148 milligrams per kilogram, indicating that substantial exposure reduction would be needed across the entire allotment complex. The authors suggest measures including targeted removal of the most contaminated soil, clean-soil replacement or capping, raised cultivation beds, dust suppression, and limiting children’s direct soil contact.

The researchers are careful to frame their conclusions within the study’s limitations. The 36 samples may not capture the full spatial heterogeneity of the allotments, plot-level differences and hotspots were not modeled, and seasonal variation in soil moisture, redox conditions, and gardening activity was not assessed. Total arsenic measurements cannot distinguish arsenic oxidation states or mineral forms, so the low bioaccessibility remains a plausible but not directly confirmed explanation rooted in mineralogy. The assessment also excluded vegetable consumption, water exposure, and co-exposure to other mining-related contaminants, meaning the true cumulative burden on residents could be higher. Nevertheless, the work demonstrates the value of combining paired bioaccessibility measurements with probabilistic simulation, offering regulators a more realistic — and more defensible — basis for deciding how to protect the gardeners of Złoty Stok, and by extension the many communities worldwide living on the contaminated soils of former mining regions.

Subject of Research: Probabilistic human health risk assessment of arsenic-contaminated allotment garden soils in a former mining town in Poland using Monte Carlo simulation and bioaccessibility testing

Article Title: Human health risk and remedial goal assessment using Monte Carlo simulation: a case study of arsenic contaminated allotment gardens in southern Poland

Article References: Olabode, K., Lewińska, K., & Komorowicz, I. (2026). Human health risk and remedial goal assessment using Monte Carlo simulation: a case study of arsenic contaminated allotment gardens in southern Poland. Environmental Geochemistry and Health, 48(14), Article 571. https://doi.org/10.1007/s10653-026-03462-8

Image Credits: AI Generated

DOI: 10.1007/s10653-026-03462-8

Keywords: arsenic, soil contamination, Monte Carlo simulation, bioaccessibility, allotment gardens, human health risk assessment, mining legacy, remedial goals, cancer risk, Złoty Stok, Poland, exposure pathways