Micronutrient deficiencies (MNDs) remain widespread among people in sub-Saharan Africa^1–5, where access to sufficient food from plant and animal sources that is rich in micronutrients (vitamins and minerals) is limited due to socioeconomic and geographical reasons^4–6. Here we report the micronutrient composition (calcium, iron, selenium and zinc) of staple cereal grains for most of the cereal production areas in Ethiopia and Malawi. We show that there is geospatial variation in the composition of micronutrients that is nutritionally important at subnational scales. Soil and environmental covariates of grain micronutrient concentrations included soil pH, soil organic matter, temperature, rainfall and topography, which were specific to micronutrient and crop type. For rural households consuming locally sourced food—including many smallholder farming communities—the location of residence can be the largest influencing factor in determining the dietary intake of micronutrients from cereals. Positive relationships between the concentration of selenium in grain and biomarkers of selenium dietary status occur in both countries. Surveillance of MNDs on the basis of biomarkers of status and dietary intakes from national- and regional-scale food-composition data^1–7 could be improved using subnational data on the composition of grain micronutrients. Beyond dietary diversification, interventions to alleviate MNDs, such as food fortification^8,9 and biofortification to increase the micronutrient concentrations in crops^10,11, should account for geographical effects that can be larger in magnitude than intervention outcomes.

Selenium (Se) is an essential element for human health and livestock productivity. Globally, human Se status is highly variable, mainly due to the influence of soil types on the Se content of crops, suggesting the need to identify areas of deficiency to design targeted interventions. In sub-Saharan Africa, including Ethiopia, data on population Se status are largely unavailable, although previous studies indicated the potential for widespread Se deficiency. Serum Se concentration of a nationally representative sample of the Ethiopian population was determined, and these observed values were combined with a spatial statistical model to predict and map the Se status of populations across the country. The study used archived serum samples (n = 3269) from the 2015 Ethiopian National Micronutrient Survey (ENMS). The ENMS was a cross-sectional survey of young and school-age children, women and men. Serum Se concentration was measured using inductively coupled plasma mass spectrometry (ICPMS). The national median (Q1, Q3) serum Se concentration was 87.7 (56.7, 123.0) μg L⁻¹. Serum Se concentration differed between regions, ranging from a median (Q1, Q3) of 54.6 (43.1, 66.3) µg L⁻¹ in the Benishangul-Gumuz Region to 122.0 (105, 141) µg L⁻¹ in the Southern Nations, Nationalities, and Peoples’ Region and the Afar Region. Overall, 35.5% of the population were Se deficient, defined as serum Se < 70 µg L⁻¹. A geostatistical analysis showed that there was marked spatial dependence in Se status, with serum concentrations greatest among those living in North-East and Eastern Ethiopia and along the Rift Valley, while serum Se concentrations were lower among those living in North-West and Western Ethiopia. Selenium deficiency in Ethiopia is widespread, but the risk of Se deficiency is highly spatially dependent. Policies to enhance Se nutrition should target populations in North-West and Western Ethiopia.

Grain and soil were sampled across a large part of Amhara, Ethiopia in a study motivated by prior evidence of selenium (Se) deficiency in the Region's population. The grain samples (teff, Eragrostis tef, and wheat, Triticum aestivum) were analysed for concentration of Se and the soils were analysed for various properties, including Se concentration measured in different extractants. Predictive models for concentration of Se in the respective grains were developed, and the predicted values, along with observed concentrations in the two grains were represented by a multivariate linear mixed model in which selected covariates, derived from remote sensor observations and a digital elevation model, were included as fixed effects. In all modelling steps the selection of predictors was done using false discovery rate control, to avoid over-fitting, and using an a-investment procedure to maximize the statistical power to detect significant relationships by ordering the tests in a sequence based on scientific understanding of the underlying processes likely to control Se concentration in grain. Cross-validation indicated that uncertainties in the empirical best linear unbiased predictions of the Se concentration in both grains were well-characterized by the prediction error variances obtained from the model. The predictions were displayed as maps, and their uncertainty was characterized by computing the probability that the true concentration of Se in grain would be such that a standard serving would not provide the recommended daily allowance of Se. The spatial variation of grain Se was substantial, concentrations in wheat and teff differed but showed the same broad spatial pattern. Such information could be used to target effective interventions to address Se deficiency, and the general procedure used for mapping could be applied to other micronutrients and crops in similar settings.

The dataset comprises primary data for the concentration of 29 mineral micronutrients in cereal grains and up to 84 soil chemistry properties from GeoNutrition project surveys in Ethiopia and Malawi. The work provided insights on geospatial variation in the micronutrient concentration in staple crops, and the potential influencing soil factors. In Ethiopia, sampling was conducted in Amhara, Oromia, and Tigray regions, during the late-2017 and late-2018 harvest seasons. In Malawi, national-scale sampling was conducted during the April–June 2018 harvest season. The concentrations of micronutrients in grain were measured using inductively coupled plasma mass spectrometry (ICP-MS). Soil chemistry properties reported include soil pH; total soil nitrogen; total soil carbon (C); soil organic C; effective cation exchange capacity and exchangeable cations; a three-step sequential extraction scheme for the fractionation of sulfur and selenium; available phosphate; diethylenetriaminepentaacetic acid (DTPA)-extractable trace elements; extractable trace elements using 0.01 M Ca(NO₃)₂ and 0.01 M CaCl₂; and isotopically exchangeable Zn. These data are reported here according to FAIR data principles to enable users to further explore agriculture-nutrition linkages.

A two-year study was conducted to assess how mulch influences weed dynamics following imposition of different fertilization treatments under three crop establishment options: (i) conventional; (ii) ripping; and (iii) basin, in a two-year maize-legume rotation. Eight treatments were imposed within each crop establishment option and received maize stover mulch applied at 0% or 30% cover before planting maize (Zea mays) or cowpea (Vigna unguiculata) as test crops. Maize received nitrogen (N) at 35, 90, or 120 kg·ha⁻¹ and phosphorus (P) at 14 or 26 kg·ha⁻¹ applied alone or in combination with 4 or 7 t cattle manure·ha⁻¹, while cowpea received 8 or 17 N·kg·ha⁻¹ and similar P rates to maize. Results indicated that both weed biomass and diversity were influenced more by fertilization than method of crop establishment. On treatments under high fertilizer application rates, or previously planted to cowpea weed biomass ranged between 220 and 400 g·m⁻² under mulch and 370–510 g·m⁻² (no mulch). Here species richness ranged between 7–16 and was dominated by dicotyledons. This was in contrast to biomass ranges of 75–200 g·m⁻² in the low fertilized and control plots, where only one or two grass types dominated. Overall, weed densities were 6% to 51% higher under conventional tillage compared to the two conservation agriculture (CA) options, although the data indicated that mulch significantly (p < 0.05) depressed weed density by up to 70%. We concluded that mulching could be a potential mechanism for reducing weeding labor costs for smallholders and the general environmental and health concerns associated with the use of herbicides in CA systems.