From reservoir to plate: monitoring irrigation water source as the dominant driver of vegetable microbial contamination risk in Sahelian smallholder systems.

Consumption of raw vegetables is a major foodborne exposure pathway in sub-Saharan Africa, yet the relative contribution of irrigation water quality versus on-farm hygiene practices to human contamination risk has rarely been quantified using a convergent, multi-method statistical framework. This study determined the relative weight of irrigation water source and Good Agricultural/Hygiene Practices (GAP/GHP) compliance as drivers of microbiological contamination, and thus human exposure risk, for raw vegetables in Northern Burkina Faso. Lettuce (n = 66) and tomato (n = 54) were sampled from 120 farms across three sites irrigated with reservoir, well, or borehole water. GAP/GHP compliance was scored by field audit; Escherichia coli and Salmonella spp. were quantified by ISO reference methods and analysed using ANOVA, logistic regression, random forest modelling, principal component analysis and k-means clustering. Irrigation water source was the dominant predictor of E. coli load (F = 61.6, p < 0.001, eta-squared = 0.51), with mean loads of 3.00, 2.25 and 1.19 log CFU/g for reservoir, well and borehole water respectively; no Salmonella was detected in borehole-irrigated samples. Reservoir water was the primary Salmonella risk factor (OR = 3.82, 95% CI 1.29-11.3). Random forest modelling confirmed water source as the leading predictor (41.1% importance) ahead of hygiene practices (9.8%), and this ranking was stable across five independent analytical methods. We translate these findings into a field-deployable Water-Security Risk Score for extension services, and situate irrigation water source as a broadly relevant human exposure pathway across smallholder vegetable systems in the wider Sahel.