Fertilizer Use Efficiency: Where Unused N, P and K Go
Fertilizer use efficiency is lower than most farmers assume. In the season of application, cereal crops typically take up 35–45% of the nitrogen (N) applied, 10–25% of the phosphorus (P) and about 40–50% of the potassium (K). But the unrecovered share behaves differently for each nutrient. Most unrecovered N leaves the field. Most unrecovered P stays in the soil and feeds later crops.
That difference decides what to change. For nitrogen, the problem is loss, and rate and timing do most of the work. For phosphorus, the problem is access, and soil P status and placement matter more than the product.
How much fertilizer does a crop actually take up?
A cereal crop takes up about 39% of fertilizer N in the season it is applied, according to a 2022 global meta-analysis by Yu and colleagues in Agriculture, Ecosystems & Environment (3,586 field observations). The study estimated 42% for rice and wheat and 36% for maize. Farmer fields often do worse: a 2010 review in Ambio reported on-farm N recovery of only 16–18% in the wheat–maize system of the North China Plain.
Phosphorus looks worse at first. FAO Fertilizer and Plant Nutrition Bulletin 18 (Syers, Johnston and Curtin, 2008) reported first-season recovery of 10–25%. When P removed by crops was compared with P applied over many years, including in Rothamsted's long-term experiments, recovery reached 50–90%. For potassium, Dobermann's 2007 review for the International Fertilizer Association summarized recovery in Asian field trials at 38–51%.
| Nutrient | First-season recovery | Long-term recovery | Where the rest goes |
|---|---|---|---|
| Nitrogen | 36–42% in trials; 16–18% in over-fertilized fields | Low | Air and water; some to soil organic matter |
| Phosphorus | 10–25% | 50–90% | Bound to soil minerals, mostly available later |
| Potassium | 38–51% | High on clays, lower on sands | Held in clay layers; leached from sands |
Climate changes these numbers. The 2022 meta-analysis found N recovery rose with growing-season temperature for rice and wheat, but fell for maize.
Where does the rest of the nitrogen go?
Unrecovered nitrogen leaves through ammonia volatilization, nitrate leaching and denitrification, or stays temporarily in soil organic matter. Which route dominates depends on the product, its placement and the weather in the first two weeks after application.
Volatilization. Urea left on the soil surface converts to ammonia gas. A 2017 meta-analysis by Silva and colleagues in Agronomy Journal found surface urea lost an average of 31% of its N this way. Losses are highest on warm, moist soils under crop residue, such as no-till maize fields in southern Brazil.
Leaching. Nitrate moves with drainage water below the root zone. The IPCC 2019 Refinement uses a default of 24% of applied N lost to leaching and runoff where water drains through the soil, and zero in dry regions.
Denitrification. In waterlogged soil, microbes convert nitrate into gases. The IPCC 2019 Refinement sets direct nitrous oxide (N₂O) emissions from synthetic fertilizer at 1.6% of applied N in wet climates and 0.5% in dry climates. Most denitrified N leaves as N₂, which is hard to measure and can be large in flooded rice.
Soil storage. Part of the "missing" N is held in soil organic matter. It returns slowly and cannot be timed to crop demand.
These routes draw from the same pool. If you stop volatilization by burying urea, more ammonium stays in the soil, where it can later leach or denitrify.
Phosphorus and potassium: fixed is not the same as lost
Phosphorus fixation means P binds to iron, aluminium and calcium compounds in the soil. It does not disappear. Johnston and Syers argued in Better Crops (2009) that the common "10–20% P efficiency" figure is mostly a result of the calculation method. Real P losses come from erosion and runoff, which are small in mass but serious for water quality.
The exception is highly weathered tropical soils, such as the Ferralsols of the Brazilian Cerrado and western Kenya. There, much applied P stays unavailable for years, and placing P close to roots matters most.
Potassium is held between the layers of clays such as illite and released slowly. On sandy soils with low cation exchange capacity (CEC), K moves with drainage water. A 2025 study by Nigon and colleagues in Soil Science Society of America Journal found sandy soils retained little K, especially when calcium and magnesium competed for exchange sites. Liming acid sandy soils raises CEC and reduces this loss.
Which lever raises fertilizer use efficiency most?
For nitrogen, rate and timing usually raise fertilizer use efficiency more than product choice, because they reduce how much N is exposed to loss at all. Source and placement work best when they target the loss route that dominates on a specific field.
Rate
Recovery efficiency falls as the rate rises (Dobermann, 2007), so where over-application is common, cutting the rate gives the largest single gain. A 2018 Nature study by Cui and colleagues followed 20.9 million Chinese smallholders: improved practices cut N use by 14.7–18.1% while yields rose 10.8–11.5%. A 2025 study in Nature Sustainability found 55% of rice farmers in the South Asian region studied over-applied N and could save 18 kg N/ha without yield loss.
Timing
Splitting N to match crop demand reduces the N sitting in the soil while roots are small. A 2020 field study in the US upper Midwest (PLOS ONE) found a three-way split raised maize recovery efficiency by an average of 30% compared with fall or single pre-plant application.
Source
Enhanced-efficiency fertilizers help only when their target loss is present. NBPT, a urease inhibitor, reduced volatilization from 31% to 14.8% of applied N in the 2017 Agronomy Journal meta-analysis. Across urease and nitrification inhibitors, a 2014 meta-analysis by Abalos and colleagues found average gains of 7.5% in yield and 12.9% in N use efficiency.
Placement
Urea deep placement, developed by IRRI and IFDC for flooded rice, puts 1–3 g urea briquettes 7–10 cm deep; IFDC reports urea savings of 25–40% and yield gains of 15–20% in Bangladesh. For phosphorus, a 2021 meta-analysis of 39 studies in Soil & Tillage Research found deep-band placement gave the clearest yield benefit in wheat and maize. On well-supplied soils the gain is small: Iowa State University Extension trials found P placement seldom changed corn or soybean yield.
Why the levers are not additive
Cutting the rate reduces every loss route at once. Splitting mainly reduces leaching and denitrification. NBPT and incorporation both target volatilization, so using both gives little more than using one. Identify the dominant loss route first, then choose one lever for it.
Worked example: maize on one hectare
A farmer broadcasts 120 kg N/ha (107 lb N/acre) as surface urea at planting: 261 kg of urea per hectare (233 lb/acre). At an assumed farm-gate price of US$0.70 per kg (use your local price), this costs US$183/ha. At 33% recovery, the crop takes up 40 kg of fertilizer N, so each kilogram reaching the crop costs about US$4.60.
Now split the same urea: one-third incorporated at planting, two-thirds side-dressed and covered with soil at the 6–8 leaf stage (V6–V8). If recovery rises by the 30% from the upper Midwest study, to about 43%, the crop needs only 93 kg N/ha (83 lb/acre) to take up the same 40 kg. That is 202 kg of urea (180 lb/acre), costing US$141 — a saving of US$42/ha before the extra field pass, assuming yield is unchanged.
Adding NBPT here would add little: incorporation already removed most volatilization risk. Spend that money on a strip trial to confirm the rate.
Start with the loss route, not the product
Sandy soil in a wet season points to split timing. Surface urea on warm, residue-covered soil points to incorporation. Highly weathered, low-P soil points to banding P. Valora Earth's WhatsApp advisor combines a field's soil type, crop stage and local rainfall forecast to suggest when to split and where to place fertilizer, in the farmer's own language.
Frequently asked questions
What is a good nitrogen use efficiency for maize?
In the season of application, maize typically recovers about 36% of fertilizer N, according to a 2022 global meta-analysis. Well-managed fields with split applications and rates matched to realistic yield often reach 50% or more. Recovery below 30% usually signals over-application, poor timing or large volatilization losses from surface urea.
Is phosphorus fixation a permanent loss?
No. Most fixed phosphorus stays in the soil and becomes available to later crops. Long-term balances summarized by FAO in 2008 show 50–90% of applied P recovered over many seasons. Highly weathered tropical soils hold P more tightly, so placement near roots matters most there. Cutting P below crop removal slowly depletes soil reserves.
Do nitrification inhibitors pay back?
They pay back where leaching or denitrification losses are high: wet climates, sandy soils, or N applied well before the crop needs it. A 2014 meta-analysis found average gains of 7.5% in yield and 12.9% in N use efficiency. On dry soils with split applications, the gain is often too small to cover the added cost.
