**Background:** Malnutrition is widespread in sub-Saharan Africa, and in rural Northern Ghana smallholder households depend on both own agricultural production and market purchases for their dietary needs. Seasonality is a key driver of undernutrition, with the dry season (hunger period) limiting availability of nutrient-dense foods like vegetables and fruits. This study aimed to model which crops should be grown, the minimum farm size required, and the potential contribution of mainstream agricultural interventions to provide a nutritious diet year-round in Northern Ghana, either through own production or income generation.
**Methods:** The study was conducted in Karaga sub-district, Northern Region, Ghana, where 32% of children under 5 were stunted and 9% wasted in 2014. Dietary intake data were collected from 337 children aged 6–23 months using quantitative multi-pass 24-hour recalls in July 2014 (hunger season). An optimal nutrient-adequate diet for non-breastfed children aged 12–23 months was developed using linear programming (e-Optifood), considering 13 key nutrients. This diet was scaled to an average household (14 members) and included 206 kg whole grains, 92.4 kg beans, 69 kg nuts/seeds, 67.9 kg dark green leafy vegetables, 61.2 kg vitamin C-rich vegetables, 262.7 kg other fruits, 64.6 kg eggs, and other foods annually. The year was divided into four seasons: Season 1 (Nov–Jan, dry, no deficits), Season 2 (Feb–Apr, dry, deficits), Season 3 (May–Jul, rainy, deficits), Season 4 (Aug–Oct, rainy, no deficits). Two scenarios were modeled using GAMS: Scenario A prioritized own production to meet food needs; Scenario B allowed all produce to be sold and all food purchased. Interventions included: (1) no intervention (average yields), (2) expanding availability via storage or irrigation, and (3) improved yields (best attainable or 50% above average) for grains, starchy roots, legumes, and vegetables.
**Key Results:** With average yields and priority on own production (Scenario A), a minimum farm size of 1.43 ha was needed, determined by Season 4 requirements. Increasing yields reduced this: best attainable legume yields required 1.00 ha, and best grain yields required 1.02 ha. Vegetable yield increases minimally reduced farm size to 1.42 ha. Without irrigation, vegetable and fruit needs could not be met by own production in Season 2. Expanding availability via storage only partly closed this gap (extra cost 50 GH₵ vs. 200 GH₵ for purchasing). Irrigation allowed full coverage and more than doubled revenue (227% of baseline in GH₵/year/ha). Improving vegetable yields increased revenue by 142% in Scenario A and 147% in Scenario B. In Scenario B (all food purchased), only 0.10 ha cultivated with onions was sufficient to earn the 9,900 GH₵/year needed for the optimal diet. Improving yields of grains, starchy roots, and legumes did not increase revenue per hectare but reduced land needed. The most lucrative crops across scenarios were onions, sweet potatoes, and watermelon.
**Clinical Implications:** The study demonstrates that average farm sizes in rural Northern Ghana (median 2.1 ha) are theoretically sufficient for year-round nutritious diets, but seasonality creates critical gaps in vegetable and fruit availability during the hunger season. Increasing yields of staple crops (especially legumes and grains) can free up land for diverse crop production, benefiting households with smaller farms (<1.43 ha). However, without irrigation or well-functioning markets, households cannot meet their vegetable and fruit needs from own production during the dry season. The findings support using food-based dietary guidelines to guide agricultural investments and highlight the need for interventions addressing both availability and affordability of nutrient-dense foods year-round. Limitations include reliance on secondary data for yields, prices, and seasonal availability, and exclusion of labor and input costs, which likely overestimate revenues.