The core argument rests on a deceptively simple geochemical fact: soil is not an inert container. Minerals, organic matter, water, plant roots, and microorganisms continuously modify the chemical forms in which contaminants exist. An identical total concentration of a metal can coexist with completely different pools of soluble, sorbed, precipitated, or organism-accessible material depending on the soil’s pH, redox conditions, dissolved organic matter, moisture, mineralogy, and particle properties. For cationic metals such as cadmium and zinc, acidification often increases solubility by weakening sorption and dissolving reactive mineral phases. Yet metalloids defy simple rules: arsenic behavior depends on its oxidation state, its binding to oxides, and competing ions. Reductive dissolution of iron oxides can liberate arsenic, while other reducing conditions may lock it away in sulfide phases. The consequence is profound: a site can host alarming total concentrations with negligible exposure, or modest concentrations with serious biological availability.

The authors draw on striking field surveys to show why occurrence data alone cannot settle questions of harm. One study detected pesticide residues in 83% of 317 European agricultural topsoils, with two or more residues in 58% of samples. A more recent survey of 63 pesticides across 373 European sites found residues at 70% of sites and identified associations with soil biodiversity and functional attributes. In greenhouse vegetable soils, total concentrations of 26 antibiotic residues ranged from 83.24 to 4,237.93 micrograms per kilogram, alongside microbial community shifts along residue gradients. These findings establish that contamination is widespread and warrant biological investigation, but the authors are careful to stress what they do not prove. A detection is not an adverse-effect threshold, and an observational association between residues and biodiversity does not establish causality, a concentration threshold, or demonstrated human-health risk.

Central to the paper is a set of distinctions that are routinely blurred in practice. The authors distinguish a potentially accessible contaminant pool from bioaccessibility, meaning release from the soil matrix under a defined test or biological setting, and from bioavailability, which involves actual absorption by a defined receptor. In the oral-exposure context, bioaccessibility concerns release into gastrointestinal fluids; an in vitro bioaccessible fraction predicts oral bioavailability only where that relationship has been validated. There is no universally measurable soil property called accessibility, no single extraction fraction that represents all organisms, and no assay that can be applied interchangeably across metals, organic chemicals, and particles. For particulate contaminants such as microplastics, polymer type, particle size, shape, surface weathering, and incorporation into aggregates all influence organism contact and soil structure, and a dissolved-fraction assay simply cannot capture particle-mediated effects.

Organic soil amendments provide a vivid illustration of how competing mechanisms complicate easy answers. Compost, manure-derived materials, and biochar are widely promoted as green remediation tools, but they simultaneously alter pH, sorption capacity, microbial substrates, and dissolved organic matter. Solid organic matter may retain contaminants, while soluble ligands can mobilize them. One cited study demonstrated dissolved-organic-matter-dependent mobilization and complexation of zinc, cadmium, and lead in rhizosphere soil solutions. A measured dissolved concentration therefore does not necessarily equal free-ion activity, uptake, or toxicity, and amendment composition, dose, and ageing all require evaluation alongside pore-water chemistry and actual receptor responses. Even a successful pot experiment, in which modified biochar reduced metal uptake by ryegrass, does not establish long-term field safety.

The same rigor applies to interpreting biological measurements. Biomarkers can provide early warning without demonstrating population decline, community composition may shift while some process rates remain stable, and unchanged species richness can conceal the loss of sensitive taxa or specialized functions. The authors argue that ecological assessment should combine organism tests with direct process measurements, targeting functional endpoints such as decomposition, nitrogen transformation, soil respiration, aggregate stability, and plant productivity. Interpretation demands matched reference soils and adjustment for moisture, pH, substrate supply, and land management. Increased respiration after adding compost may merely reflect the new carbon substrate rather than recovery, while reduced respiration could signal stress or simply altered substrate availability. Cited research showed that soil pH moderated the resistance and resilience of carbon- and nitrogen-cycling functions under stress, underlining how soil state shapes functional outcomes. Nor should synergy or additive effects in contaminant mixtures be assumed; testing the actual contaminated soil or mixture is preferable where feasible.

The familiar screening tools retain their roles but must be read correctly. The geoaccumulation index compares a concentration with adjusted background values; the enrichment factor normalizes to a reference element; both indicate enrichment, not biological harm, and both depend critically on appropriate background and reference choices. The potential ecological risk index combines contamination factors with toxic-response weights, but its original sediment-based formulation should not be transferred to soils or new contaminant classes without justification. Hazard quotients, whether comparing estimated human oral dose with a reference dose or exposure with an ecological benchmark, identify potential concern when they exceed one, but a quotient is not a probability of harm, and values below one do not resolve unassessed pathways or substances. Crucially, the authors warn that bioaccessibility must never be inserted as an arbitrary multiplier into these calculations; any adjustment requires receptor-specific validation and applicable assessment guidance.

To operationalize the approach, the Perspective lays out an eight-step workflow connecting problem formulation, baseline characterization, screening and prioritization, receptor-specific exposure refinement, functional assessment, risk characterization, intervention selection, and verification. Problem formulation means documenting sources, land-use history, migration pathways, and receptors, and explicitly specifying which soil functions to protect. Baseline work demands spatially representative sampling across seasons and depths, with matched reference soils and full quality controls. Exposure refinement selects methods suited to the contaminant and pathway: pore-water measurements, passive sampling, speciation, validated bioaccessibility assays, uptake measurements, or particle characterization. Intervention selection then compares technologies, from source removal and soil washing to thermal treatment, immobilization, microbial degradation, and plant-assisted remediation, against exposure reduction, durability, secondary wastes, and cross-media transfer, remembering that soil washing produces contaminated fractions that still require management and immobilization leaves the contaminant inventory in place.

The framework’s proposed decision rule is perhaps its most consequential element: a remediated site should demonstrate acceptable residual exposure, recovery of the specified soil functions, and control of contaminant transfer, confirmed through repeated observations across relevant seasons and plausible disturbances such as flooding or drying. Rebound, persistent impairment, or new transfer pathways should trigger renewed investigation. The authors also embrace modern predictive tools with appropriate skepticism: suspect screening, environmental DNA, and omics can characterize composition and functional potential, but gene presence does not establish current process activity, and spatial models such as SoilGrids 2.0 illustrate uncertainty-aware prediction rather than direct validation of contaminant-risk forecasts. Notably, the authors frame their own contribution honestly, stating that the added predictive value of this integration remains a testable proposition rather than an established universal result.

If the framework performs as hypothesized, the implications reach far beyond the scientific literature. Multi-site comparisons are needed to test whether combining receptor-relevant exposure, soil-state variables, and functional indicators genuinely improves prediction of impairment and recovery relative to total concentrations alone, and the authors acknowledge limitations including laboratory-to-field extrapolation, confounding in observational studies, assay comparability, mixture uncertainty, and resource constraints. Yet the direction of travel is clear. As pesticide residues, antibiotics, metals, and microplastics accumulate in soils worldwide, the question that increasingly matters is not simply how much contaminant is in the ground, but whether the soil can still decompose litter, cycle nitrogen, aggregate, and support plants, and whether a cleanup restores those capacities durably. Measuring what soil does, rather than merely what it contains, may be the shift that turns contamination monitoring into genuine environmental health assessment.

Subject of Research: Function-centered risk assessment for soil contaminants linking exposure, bioavailability, and ecosystem function

Article Title: Soil contaminants at the environment-health interface require function-centered risk assessment

Article References: Izah, S. C., Aseibai, E. R., Promise, V. I., & Ogwu, M. C. (2026). Soil contaminants at the environment-health interface require function-centered risk assessment. Frontiers in Soil Science, 6, Article 1954312. https://doi.org/10.3389/fsoil.2026.1954312

Image Credits: AI Generated

DOI: 10.3389/fsoil.2026.1954312

Keywords: soil contamination, risk assessment, bioavailability, bioaccessibility, soil functions, remediation, heavy metals, pesticides, microplastics, ecotoxicology, ecosystem processes, environmental health