The research team, led by Omolara Jemimah Ojezele of the Federal College of Animal Health and Production Technology in Ibadan and Matthew Obaineh Ojezele of Delta State University, focused on Amaranthus cruentus, a leafy green vegetable that is a dietary staple across Nigeria and much of sub-Saharan Africa. Sampling was carried out at 30 sites spread across three major vegetable-producing settlements in Oyo State, yielding 240 matrix-season samples of water, sediment, soil, and vegetable tissue. After analytical duplicates were averaged, the dataset comprised 480 independent pesticide-specific sample-level observations. All samples were analysed using validated chromatographic methods, providing the analytical foundation for the layered statistical and risk assessments that followed.

The study’s first major finding concerns environmental partitioning, the tendency of a chemical to accumulate in particular compartments of an ecosystem. Glyphosate, a systemic herbicide that binds strongly to soil particles, predominated in soil and sediment samples, consistent with its known physicochemical behaviour. Cypermethrin, a synthetic pyrethroid insecticide, showed a different pattern, occurring at comparatively higher concentrations in the edible Amaranthus cruentus tissue itself. This divergence matters because it shapes the primary route of human exposure: while glyphosate residues largely remain locked in the ground, cypermethrin residues travel directly onto the plates of consumers through the vegetable crop, bypassing many of the natural attenuation processes that limit exposure to soil-bound chemicals.

To understand whether contamination was random or clustered, the researchers applied Moran’s I spatial autocorrelation analysis, a statistical technique that tests whether values measured at nearby locations are more similar than expected by chance. The results were striking for one compound: glyphosate in sediment displayed significant spatial autocorrelation, with a Moran’s I statistic of 0.783 and a p-value of 0.0003. In practical terms, this means that sites with high glyphosate contamination in sediment tended to sit near other highly contaminated sites, forming identifiable hotspots rather than a scattered pattern. Such clustering suggests that localised factors, such as application intensity, topography, and runoff pathways, drive glyphosate accumulation in sediments, information that could help regulators target interventions at specific zones rather than imposing blanket restrictions.

The team complemented the spatial analysis with hierarchical mixed-effects modelling and principal component analysis, statistical tools that allowed them to separate the influence of different environmental and seasonal variables on residue concentrations while accounting for the nested structure of the sampling design. Together with a structured survey of vegetable farmers, these methods revealed a troubling gap between residue levels in crops and the practices that produced them. The survey documented limited pesticide stewardship among growers, including short pre-harvest intervals, meaning farmers were applying pesticides and harvesting their crops with little time for residues to degrade. Substantial exceedances of selected European Union maximum residue limit benchmarks for spinach were recorded in the vegetable samples, underscoring the consequences of these practices.

The heart of the study lies in its dietary risk assessment, which the researchers conducted in two complementary ways. The first was a deterministic approach, which uses single point estimates of residue concentration and consumption to calculate a hazard quotient, the ratio of estimated exposure to a reference dose below which no adverse effects are expected. A hazard quotient above one signals a potential health concern. The deterministic results indicated acceptable dietary risk for adults for both pesticides, and acceptable risk for children for glyphosate. However, cypermethrin exceeded the hazard threshold in children, with a hazard quotient of 1.818, nearly double the level of concern for the youngest and most vulnerable consumers.

Because deterministic assessments rely on single values, they can obscure the true range of possible exposures. To address this, the researchers turned to Monte Carlo simulation, a probabilistic technique that runs thousands of virtual exposure scenarios by randomly drawing residue concentrations and consumption rates from their measured distributions. Each of the 10,000 iterations represents one plausible combination of real-world conditions, and the resulting spread of hazard quotients reveals how often risky exposures might actually occur. The simulation results sharpened the study’s central warning: hazard quotients above one appeared in 15.09 percent of adult cypermethrin exposure scenarios and a remarkable 70.47 percent of child exposure scenarios. For glyphosate, by contrast, no simulated exposure scenario for either group exceeded the threshold, suggesting the herbicide poses comparatively limited dietary risk under the conditions measured.

The contrast between the two pesticides carries important implications. Cypermethrin is prized by farmers for its effectiveness against the insect pests that devastate vegetable crops, but its persistence on leafy surfaces means that harvest timing is critical to food safety. The finding that roughly seven in ten simulated child exposure scenarios exceeded the hazard threshold points to a systemic problem rather than isolated lapses, one rooted in the short intervals between spraying and harvest documented in the farmer survey. Children are particularly sensitive in such assessments because their lower body weight means the same residue concentration in food delivers a higher dose per kilogram of body mass, and because developing systems may be more vulnerable to neurotoxic compounds such as pyrethroids.

The study’s authors argue that their findings support a coordinated response spanning pesticide stewardship, residue surveillance, farmer education, and regulatory oversight. In many smallholder farming systems across sub-Saharan Africa, extension services that once trained farmers in safe application practices have contracted, leaving growers to rely on agrochemical vendors and informal knowledge networks. The short pre-harvest intervals recorded in this study are a classic symptom of that gap, reflecting neither malice nor ignorance of risk so much as the absence of accessible, practical guidance. Simple interventions, such as clear labelling in local languages, community-level training on waiting periods, and routine market testing of leafy vegetables, could substantially reduce exposure without requiring farmers to abandon chemical pest control altogether.

The research also demonstrates the value of methodological rigour in regions where pesticide monitoring data have historically been sparse. By pairing validated chromatographic analysis with spatial statistics, mixed-effects modelling, and Monte Carlo simulation, the study transforms a set of field measurements into a nuanced risk picture that distinguishes between compounds, environmental compartments, and consumer groups. The strong spatial clustering of glyphosate in sediment, for instance, would have been invisible to a conventional point-sampling survey, and the probabilistic approach revealed a child-specific risk that a deterministic calculation alone might have understated or overstated. As pesticide use intensifies across African agricultural landscapes to meet growing food demand, studies of this kind provide the evidence base that national regulators and international bodies such as the Joint FAO/WHO Meeting on Pesticide Residues need to set meaningful, locally calibrated safety standards. For now, the message from Oyo State is clear: the greens on the market stall are nutritious and affordable, but for the children eating them, the invisible residues of an everyday insecticide may be crossing a line that better farming practice could hold firmly in place.

Subject of Research: Pesticide residue occurrence and dietary health risk assessment in agricultural ecosystems of southwestern Nigeria

Article Title: Environmental occurrence, spatial distribution, and probabilistic dietary risk assessment of cypermethrin and glyphosate residues in agricultural ecosystems of southwestern Nigeria

Article References: Ojezele, O. J., & Ojezele, M. O. (2026). Environmental occurrence, spatial distribution, and probabilistic dietary risk assessment of cypermethrin and glyphosate residues in agricultural ecosystems of southwestern Nigeria. Environmental Monitoring and Assessment, 198(11), Article 1171. https://doi.org/10.1007/s10661-026-15973-z

Image Credits: AI Generated

DOI: 10.1007/s10661-026-15973-z

Keywords: cypermethrin, glyphosate, pesticide residues, Amaranthus cruentus, food safety, dietary risk assessment, Monte Carlo simulation, spatial autocorrelation, Nigeria, vegetable farming, pesticide stewardship, environmental monitoring

The team installed pollen traps at two marked hives per apiary, pooling material into composite two-week samples across seven sampling periods each season, from early May to mid-August in 2025 and mid-May to late August in 2024. Samples were frozen at minus 18 degrees Celsius until analysis, and accredited laboratories in Berlin used validated multi-residue methods based on liquid and gas chromatography coupled with tandem mass spectrometry, with dedicated LC-MS/MS screening for the highly polar herbicide glyphosate and its breakdown product AMPA. Botanical composition was reconstructed under light microscopy, with more than 500 pollen grains identified per sample, allowing the researchers to link each residue detection to the plants the bees had actually visited.

In total, 15 active substances were detected in 2024 and 19 in 2025. Fungicides dominated the detection profile in both years, accounting for 51 percent of detections in 2024 and 54 percent in 2025, followed by herbicides and, trailing behind, insecticides. The most frequently recorded compounds included boscalid and azoxystrobin among the fungicides, glyphosate and clopyralid among the herbicides, and the neonicotinoid-class insecticide acetamiprid. Maximum concentrations were striking: glyphosate reached 840 micrograms per kilogram in 2024 and 327 in 2025, boscalid peaked at 610 micrograms per kilogram in 2024, and the herbicide fluazifop-P-butyl hit 660 and 1800 micrograms per kilogram across the two years, far exceeding the European Union recommended maximum residue level of 50 micrograms per kilogram for bee products.

The fluazifop-P-butyl findings stand out as among the most consequential for beekeepers and regulators. This grass-specific herbicide is not typically considered a bee-toxicity headline grabber, yet its concentrations in Latvian pollen exceeded most values reported elsewhere in Europe, though they remained below the extreme 6831 micrograms per kilogram documented in southern Germany by earlier researchers. Acetamiprid tells a different regulatory story. The European Union’s acceptable concentration in bee products was 50 micrograms per kilogram in 2024 but rose to 300 micrograms per kilogram in 2025, and individual Latvian samples exceeded the limits applicable at the time of sampling in both years, with maxima of 350 and 200 micrograms per kilogram respectively.

Perhaps the most unexpected detection was terbutryn, a herbicide banned for agricultural use in the EU, found once in 2025 exclusively at the urban-influenced Jelgava apiary. Because it appeared in no earlier sampling year, the researchers rule out current farming as the source. Terbutryn is known to leach from biocide-treated facade coatings and paints, and it is used for aquatic algae control; a wastewater treatment plant within the foraging radius of the Jelgava hives may represent a plausible contamination pathway. The finding is a reminder that pollinator pesticide exposure is not confined to fields, and that urban infrastructure can quietly contribute to the chemical loads carried back to the hive.

Seasonal patterns proved decisive. Residue concentrations and risk values rose mainly between mid-May and early July, coinciding with the flowering of oilseed rape and legumes and the associated pesticide applications. In 2024, the highest total residue load occurred in the second sampling period, when pollen from the crucifer family dominated the samples and bees were foraging intensively on winter oilseed rape. Elevated levels followed in early June, tied to legume crops such as field beans and peas, and again in late July when flowering weeds of the goosefoot family provided a late-summer pollen source. In 2025, detections were spread more evenly across the season, with peaks in late May and early June linked to crucifers and rose-family bloom such as apple, plum and pear, and renewed elevations in late July associated with summer oilseed rape.

To translate chemical concentrations into biological risk, the team applied the Pollen Hazard Quotient, a screening metric that divides the detected concentration in pollen by the acute oral or contact LD50 for honey bees, assuming a maximum daily intake of up to 9.5 milligrams of bee bread per nurse bee. Values below 50 are considered low risk, between 50 and 500 moderate, and above 500 high. Median PHQ values stayed safely under 50 in every sampling period of both years, indicating generally low colony-level toxicological pressure. But individual samples told a sharper story: several 2025 samples crossed the threshold, and one sample from Vecauce parish in the fifth period of 2025 reached a staggering PHQ of 1173.91, driven almost entirely by cypermethrin, a pyrethroid insecticide whose very low contact LD50 means even a moderate concentration translates into a high hazard score.

That cypermethrin result crystallises a central lesson of the study: toxicity, not abundance, determines risk. Frequently detected fungicides and herbicides contributed little to cumulative PHQ because their LD50 values are comparatively high, while compounds such as acetamiprid, flupyradifurone and cypermethrin punched far above their concentration weight. The researchers caution that PHQ is a screening tool with real limitations. It rests on acute endpoints only, ignoring chronic exposure, sublethal effects and the potential interactions within the pesticide mixtures that bees inevitably encounter simultaneously. Real-world exposure risk may therefore be underestimated, particularly when colonies face repeated pulses throughout a brood-rearing season.

What the study did not find may matter just as much. The team hypothesised that apiaries surrounded by more agricultural land, more nectar-producing crops, or crops under higher pesticide-use intensity would show higher cumulative PHQ values. Graphical comparisons across the six sites and two years found no consistent pattern. Platone parish, with the highest proportion of agricultural land, did not post the highest risk values, while some sites with little farmland still recorded measurable exposure. High proportions of crucifer, legume or rose pollen did not reliably track the mapped crop areas around each apiary, suggesting that bees forage selectively, that systemic pesticides move residues away from treated fields, and that broad landscape categories are simply too coarse to capture field-level application timing. Exposure, the authors conclude, is shaped by temporal dynamics, local management and bee behaviour rather than landscape composition alone.

The study builds on earlier Latvian work published in 2024 and offers the Baltic region some of its first long-term pollen monitoring data, filling a gap where oilseed rape covers 145,100 hectares and legume areas have grown 84 percent since 2020. For beekeepers, the practical message is that danger is concentrated in identifiable windows, chiefly mid-May through early July, when crop flowering and spraying overlap with intensive foraging and brood rearing. For regulators and researchers, the findings argue for monitoring designs that capture short-term exposure peaks and application timing rather than averaging the season into invisibility. Honey bee-collected pollen, the authors note, proved a suitable and temporally resolved tool for exactly this purpose, identifying precisely when and where agricultural environments turn most hazardous for their most industrious visitors.

Subject of Research: Pesticide residues and seasonal exposure risk in honey bee-collected pollen across Latvian agricultural landscapes

Article Title: Pesticide exposure risks and seasonal variation in honey bee-collected pollen across agricultural landscapes in Latvia

Article References: Pesticide exposure risks and seasonal variation in honey bee-collected pollen across agricultural landscapes in Latvia. (n.d.). https://doi.org/10.1007/s11356-026-38238-1

Image Credits: AI Generated

DOI: 10.1007/s11356-026-38238-1

Keywords: honey bees, pesticide residues, pollen, pollinator health, glyphosate, cypermethrin, Pollen Hazard Quotient, agricultural landscapes, fungicides, Latvia, ecotoxicology, bioindicators