Crop harvests for direct food use insufficient to meet the UN's food security goal

Competition among different end uses of crops—direct food, animal feed, biofuels and other industrial uses—has intensified, reducing the share of production available for direct human consumption. Yield stagnation in major cereals and rising demand from a growing global middle class for feed- and processing-intensive products raise concerns for achieving the UN Sustainable Development Goal 2 (food security by 2030). A key barrier is the lack of spatially detailed, long-term information on where crops are harvested for different initial uses. This study addresses that gap by quantifying and mapping global patterns and trends in the initial utilization of major crops, evaluating yields and nutritional outputs (calories, protein, fat) by sector, and assessing the implications for meeting caloric needs of growing and undernourished populations by 2030.

The study fuses national-level Food and Agriculture Organization Food Balance Sheets (FAO FBS; 1961–2013, pre-methodology change) with gridded global crop data to produce annual, high-resolution maps (5 arc-min) of crop harvests by initial utilization category. Seven utilization sectors are defined: food, feed, processing, export, industrial (other non-food), seed and losses. Ten major crops/product groups are included: barley, cassava, maize, oil palm (palm and palm kernel oil), rapeseed (rape and mustard seed), rice, sorghum, soybean, sugar cane and wheat, which together represent ~83% of harvested food calories and ~63% of harvested area. For each country-year and crop, FAO FBS quantities (thousand tons) are used to compute sectoral fractional utilization. These national fractions are applied to gridded harvested areas per crop to split each grid cell’s harvested area across the seven sectors and, using grid-cell yields, to derive sectoral production. Production is converted to calories, protein and fat using static crop-specific conversion factors. Aggregating across crops yields sectoral harvested area, production and yields (kcal ha⁻¹ yr⁻¹, kg protein ha⁻¹ yr⁻¹, kg fat ha⁻¹ yr⁻¹) per grid cell and globally. Global trends (1964–2013) are summarized, and linear models fitted to the most recent 20 years (1994–2013) provide projections to 2030. Model diagnostics include two-sided t-tests for slope significance (α=0.05), Lilliefors tests for residual normality and Ljung–Box Q tests for residual autocorrelation; where slopes are not significant, intercept-only models are used. Spatial maps display fractional sectoral composition by grid cell for 1964–1968 and 2009–2013 means; analogous mapping is done for projected 2030 kcal fractions. To assess SDG2 implications, the study compares additional food-sector calories projected for 2030 to country-level calorie requirements for projected population growth and undernourished populations, using 2020 ADER (average dietary energy requirement) and MDER (minimum dietary energy requirement) values. Baseline undernourishment is FAO 2018–2020; 2030 country counts are scaled from 633.98 million (2018–2020) to 656.8 million (2030) or derived from linear trends under 15 alternative scenarios. A sensitivity range of shortfalls is reported. An extreme diversion scenario also evaluates sufficiency if all harvested calories across all sectors were redirected to direct food. Uncertainties include restriction to ten crops, national-to-subnational fraction application, equating sectoral totals to adjusted production (imports, stock changes), data uncertainties (FAO, gridded yields/areas), static nutrient conversions, potential distortions in import-dependent nations, uses of exported crops in importers, and the linear projection assumption (including potential climate impacts and shifting crop locations).

• Harvested area shares shifted globally from food and feed toward processing, export and industrial uses. The fraction of harvested area for direct food fell from ~51% (1960s) to ~37% (2010s), projected ~29% by 2030; feed declined from 21% to 18%, projected 16%. Processing rose from 6% to 14%, projected 17%; export from 10% to 18%, projected 23%; industrial from 1% to 5%, projected 8% by 2030 (Table 1).
• Food-designated crops have the lowest yields (kcal, protein, fat) globally across the period; yields for industrial, export and processing sectors increased more rapidly in recent decades. Projected 2010s→2030 caloric yield increases: industrial +28%, food +24%, feed +21%; industrial caloric yields were already ~2× food and ~1.4× feed in the 2010s, widening to ~2.1× and ~1.5× by 2030.
• Spatial reconfiguration shows large regional shifts away from direct food harvests: across parts of China and India toward feed/processing/industrial; Europe toward processing/export; the United States toward processing/industrial; Brazil and Argentina toward processing/export; Malaysia and Indonesia toward export/industrial; Central Asia and parts of Russia toward export. Some African regions show reductions in food harvest fractions; overall, much of Africa exhibits limited change but low capacity for diversion.
• SDG2 sufficiency: Among 156 countries (86 with reported undernourishment), extra food-sector calories projected for 2030 will be insufficient to meet needs of the increased population plus undernourished by ~675.4 trillion kcal yr⁻¹ (MDER) or ~993.9 trillion kcal yr⁻¹ (ADER). Across 15 scenarios, global shortfalls range ~587.2–1,269.3 trillion kcal yr⁻¹ (MDER) and ~880.7–1,755.6 trillion kcal yr⁻¹ (ADER).
• Even if all harvested calories across sectors are hypothetically diverted to food, many countries—concentrated in sub-Saharan Africa—would still struggle to meet requirements due to already high shares of direct-food harvests and limited diversion potential. Of 86 countries with undernourishment studied, 31 are projected not to meet caloric requirements for both undernourished and population growth even if all harvested calories are diverted to food; an additional 17 will not meet requirements for population growth alone.

The findings demonstrate that global cropping systems have increasingly prioritized processing, export and industrial uses over direct food and feed, with higher yields and faster yield gains in non-food sectors. This allocation pattern, coupled with lower yields for food-designated crops and rising demand for animal products and processed foods, undermines progress toward SDG2 when considering calories available from crops harvested explicitly for food. Regional analyses attribute shifts to policy, market and consumption changes: biofuel and processing expansion in the United States; increased processing in Europe; commodity price and exchange-rate-driven expansion of maize and soybean in Latin America with environmental costs; dietary westernization and income-driven demand for high-value products in China; and growing processed food demand in India. The study’s sectoral maps at high spatial resolution provide tools for targeted reconfiguration of cropping systems and supply chains to enhance direct food availability. Policy implications include incentivizing production and yield growth of directly consumed food crops, and, where feasible, diverting portions of highly productive regions’ output towards food to support food-insecure nations. Equity, access and livelihood justice should guide such reallocation. The results highlight that simply increasing total production does not ensure food access and that strategic shifts in utilization and distribution are needed to address undernourishment.

This study develops the first high-resolution, global, time-series maps of initial crop utilization across seven sectors for ten major crops, quantifies sectoral yields in calories, protein and fat, and projects trends to 2030. It reveals a sustained shift of harvested area and production away from direct food and feed toward processing, export and industrial uses, with comparatively low yields for food crops. Under current trajectories, calories from crops harvested for direct food use will not suffice to eliminate undernourishment by 2030, and many countries will face shortfalls even if all harvested calories were redirected to food. The work underscores the need for policies that boost harvests and yields of directly consumed food crops and strategically redirect production to address food insecurity while considering equity and environmental sustainability. Future research should: track and model post-harvest flows between utilization categories and across borders; integrate additional crops and updated datasets; improve subnational utilization estimates; assess dynamic yield gaps over time; and optimize spatial production–utilization linkages and supply chains to meet nutritional goals.

Key limitations and uncertainties include: reliance on the top ten global crops, omitting others important locally; application of national-level utilization fractions to subnational grids due to lack of global subnational data; equating sectoral utilization totals to adjusted production (assumes proportional contributions from production, imports and stock changes), potentially distorting import-dependent countries; uncertainties in FAO FBS and gridded harvested area/yield datasets; use of static nutrient conversion factors; potential misallocation because exported crops may be used differently in importing nations or re-exported; linear projections based on 1994–2013 trends that may not capture nonlinear dynamics, climate impacts or shifting crop locations; limited country coverage for undernourishment projections and assumption that 2020 ADER/MDER values apply to 2030; and that food security involves access, utilization and stability beyond production quantities.