The impact of phosphorus on projected Sub-Saharan Africa food security futures

The study addresses how phosphorus availability and use influence Sub-Saharan Africa’s future food security under different socioeconomic trajectories. The context is the UN’s Sustainable Development Goal 2 to end hunger and achieve food security while promoting sustainable agriculture. In 2015, over half of Sub-Saharan Africa’s population was food insecure. Food security comprises availability, access, utilization, and stability. Historically, increased food demand has been met by intensification (e.g., irrigation and fertiliser), which can impose environmental costs. Phosphorus (P) is identified by FAO as the most yield-limiting nutrient with no biochemical substitute and is derived from finite phosphate rock resources. Socioeconomic pathways (SSPs) shape fertiliser use through population, GDP, policy, technology, diet, and trade. Using SSPs within the IMAGE 3.2 framework, the study examines how differing development pathways (sustainability, middle-of-the-road, regional rivalries/nationalism, inequality, and fossil-fuelled development) affect food prices, availability, hunger risk, land use, yields, and phosphorus requirements to 2050. The research question is how P requirements and management intersect with these pathways to influence food security outcomes and environmental impacts in Sub-Saharan Africa.

Background evidence highlights: (1) Phosphorus is critical to plant growth and lacks substitutes; fertiliser products (NPK, DAP, MAP, TSP) rely on phosphate rock, a finite resource. Global reserves are estimated at ~65 billion tonnes with resources ~300 billion tonnes; peak phosphorus (when high-grade supply cannot meet demand) may occur around 2070–2080. (2) Fertiliser use has risen most in developing regions; phosphate fertiliser prices (DAP, TSP) correlate strongly with phosphate rock prices (r≈0.88 since 1960) and experienced spikes in 2008 and 2011; 2010–2020 had the highest sustained prices, with another spike in 2021–2022. (3) Earlier modelling (MAGNET coupled to IMAGE) mapped SSA food security scenarios under SSPs, indicating differing population and diet trajectories. (4) Many SSA soils are highly weathered with strong P sorption to Fe/Al oxides, leading to historically low P use, nutrient depletion, and large ecological yield gaps; hysteretic crop uptake effects from legacy P seen in high-input regions are generally absent in SSA. Together, these studies motivate a coupled analysis of socioeconomic scenarios, land use, and soil chemistry to quantify P needs and food security outcomes.

The study projects SSA phosphorus requirements under SSP1–SSP5 using IMAGE 3.2 outputs for crop production, yields, land use, food prices, and socioeconomics at 0.5° × 0.5° resolution. These projections are coupled to the Geochemical Dynamic Phosphorus Pool Simulator (GDPPS), which represents soil P in two pools: a labile (partially plant-available) pool and a stable pool. Transfers between pools follow the KINS-P kinetic model derived from van der Zee & van Riemsdijk, controlled by concentrations of oxalate-extractable Fe and Al oxides (M) and an activity constant k. GDPPS mass balances include inputs (litter, fertiliser, manure, weathering, fresh soil, atmospheric deposition), outputs (runoff, crop uptake), and dissolution/precipitation between pools. Governing equations (as presented) express dL/dt and dS/dt including fluxes and kM/L coupling. Key data inputs: - IMAGE 3.2 gridded crop production and uptake demands per SSP; - Virgin pool sizes (Yang et al.) at 0.5°; - Soil Fe/Al oxides from ISRIC/Hengl et al. at 250 m; M was held constant, k and dissolution constants calibrated per grid cell using virgin soil equilibrium assumptions. Phosphorus use efficiency values (0.8–1) are inherently captured by the model structure. Scenarios: Continental projections for SSP1–SSP5 (1950–2050); country-level analyses for SSP2 (middle of the road) to derive spatial distributions of P uptake/application and national totals. Conversion and costing: Elemental P demand converted to phosphate rock (assumed 70% BPL ≈ 32% P2O5) and to DAP (46% P2O5). Minimum phosphate rock requirements assume 100% mining-to-field efficiency; a more realistic efficiency is ~80%, implying estimates are lower bounds. Cost estimates use 2010–2020 mean DAP market price (US$ 405 ± 70 per tonne) and also compare to TSP where relevant; farmgate prices are acknowledged to be 2–5× market prices due to transport, taxes, and other costs. Economic context uses published real GDP projections for selected countries for benchmarking expenditures. Model validation: GDPPS was validated previously by comparing modelled soil P pools to 17,160 geo-referenced soil profiles (1950–2000) using lack-of-fit tests, RMSE, and Willmott’s index, reported in prior open-access work.

• Under all SSPs, SSA will require large increases in fertiliser P applications to 2050. From ~560,000 tonnes of elemental P in 2016, projected 2050 annual elemental P applications are: SSP4 ≈ 1.7 million tonnes (+310%), SSP1 ≈ 2.6 million tonnes (+475%), SSP2 ≈ 3.1 million tonnes (+550%), SSP3 ≈ 3.2 million tonnes (+530%), SSP5 ≈ 3.4 million tonnes (+620%). - Food security outcomes (price, availability, risk of hunger): Only SSP1 (sustainability) and SSP5 (fossil-fuelled) improve food security by 2050, each reducing risk of hunger from ~200 million (2010) to ~50 million people. SSP2 maintains 2010 inadequacies (~200 million at risk). SSP3 and SSP4 worsen (>300 million at risk). - Economic and price dynamics by 2050 (Table 1): SSP1 shows lower relative production prices (0.59 of 2010) and higher GDP per capita (~3.21×, US$ 5300). SSP5 has slightly higher production prices (1.04) but much higher GDP per capita (~4.42×, US$ 7300). SSP3 increases production prices (1.48) with weak GDP growth (~1.21×), worsening affordability. - Production, yields, and land expansion: By 2050, highest yields in SSP1 and SSP5 (~2.45 Mg dry matter ha−1 yr−1), lowest in SSP4 (~1.94). Cropland area expands most in SSP5 (~470 Mha) and SSP3 (~460 Mha); least in SSP1 (~380 Mha). SSP3 induces high P use and cropland expansion yet worsens food security. - Minimum phosphate rock required under SSP2 (2020–2050) is ~440 million tonnes (<1% of current known global reserves of 65 billion tonnes), assuming 100% efficiency; at ~80% efficiency the true requirement would be higher. - Concentration of demand: ~74% of phosphate rock under SSP2 is required in 10 countries: Nigeria 24%, Ethiopia 12%, South Africa 7%, Ghana 7%, D.R. Congo 6%, Côte d’Ivoire 5%, Cameroon 5%, Tanzania 3%, Benin 3%, Mali 3%. Four countries (Nigeria, Ethiopia, South Africa, Ghana) account for ~50% of use. - National trajectories under SSP2: From 2010–2020 to 2050, projected increases in phosphate requirement are Nigeria +934%, Ethiopia +375%, South Africa +148%, Ghana +130%. By 2050 annual DAP-equivalent consumption is projected at ~3.19, 2.05, 0.64, and 0.92 million tonnes for Nigeria, Ethiopia, South Africa, and Ghana, respectively. - Costs: Total DAP market expenditure (2020–2050) for SSA is projected at US$ 130 ± 25 billion (farmgate potentially 2–5× higher). By 2050 annual DAP market expenditures are projected at ~US$ 1300 ± 230 million (Nigeria), US$ 815 ± 150 million (Ethiopia), US$ 250 ± 45 million (South Africa), and US$ 370 ± 70 million (Ghana). - Spatial application/uptake under SSP2: Most regions had ≤7 kg P ha−1 uptake and application in 2020; by 2050 uptake and application rise to ~12–15 kg P ha−1 in parts of West, Central, East, and Southern Africa (e.g., Benin, Ghana, Cameroon, Gabon, Ethiopia, Nigeria, Republic of Congo, South Africa, Zambia). - Hysteretic crop uptake effect: Historically present in some countries (e.g., Nigeria, South Africa, D.R. Congo, Cameroon), but under SSP2 to 2050 no country exhibits sufficient residual P to sustain increased uptake without increased inputs; all rely on fertiliser P to maintain even current food security levels.

Findings demonstrate phosphorus as a critical constraint on SSA’s food security pathways. Sustainable development (SSP1) yields improved affordability, availability, and reduced hunger risk with comparatively low P use and limited cropland expansion, achieved through higher yields and efficient, targeted P management. Fossil-fuelled development (SSP5) also improves food security but at the cost of high P inputs and substantial cropland expansion, implying greater environmental risks (e.g., runoff-driven pollution). Nationalistic regional rivalries (SSP3) expand cropland and increase P use yet worsen food security due to higher production prices and weak GDP growth, illustrating that intensification without affordability gains does not translate to food security. Inequality (SSP4) maintains lower P use but still worsens food security due to low purchasing power and limited production gains. The results emphasize that meeting SDG2 requires coupling productivity gains with equitable economic growth and efficient nutrient use to avoid environmental harm. Addressing ecological and economic yield gaps in SSA soils—with strong P sorption to Fe/Al oxides—necessitates targeted fertiliser strategies, improved land management, and consideration of diet, waste reduction, and trade within SSP-consistent policies.

The study quantifies how alternative socioeconomic pathways shape phosphorus requirements, agricultural production, and food security in Sub-Saharan Africa. Only SSP1 (sustainability) and SSP5 (fossil-fuelled) improve food security by 2050, with SSP1 achieving this with lower P inputs and minimal cropland expansion. SSP3 (nationalism) and SSP4 (inequality) worsen food security, with SSP3 additionally driving high P use and cropland expansion. Maintaining current inadequate levels of food security under SSP2 would still require at least 440 million tonnes of phosphate rock (2020–2050) and significant expenditures. Overall, sustainable economic growth with targeted, efficient phosphorus use and avoidance of nationalist trade barriers offers the most balanced route to improved food security while limiting environmental impacts.

Key caveats include: (1) Fertiliser form assumption: Costs and quantities are primarily expressed as DAP (46% P2O5); in practice, various fertilisers (e.g., NPK, TSP, MAP) are used, complicating direct substitution. (2) Costing excludes farmgate markups: Farmgate prices can be 2–5× market prices due to transport, taxes, and other costs. (3) Price bounds: Projections assume DAP prices within the 2010–2020 mean (US$ 405 ± 70 per tonne), whereas recent spikes have exceeded these levels; long-term reversion is assumed by analogy with past spikes. (4) Minimum rock requirement: The 440 Mt estimate assumes 100% mining-to-field efficiency; a more realistic ~80% efficiency implies higher actual rock demand. (5) Soil parameter stability: The Fe/Al oxide concentration (M) was held constant; while justified, real-world changes could affect sorption dynamics. (6) Scenario and model structure limitations inherent to IMAGE 3.2 and GDPPS (e.g., diet, trade, technology assumptions, gridded resolution) affect projections and their generalizability.