Global mean nitrogen recovery efficiency in croplands can be enhanced by optimal nutrient, crop and soil management practices

Nitrogen (N) is the main limiting nutrient for crop photosynthesis and growth, yet excessive N fertilizer use lowers N recovery efficiency (NUEr) and increases losses to air and water, harming ecosystems. With global food demand projected to rise by 50% by 2050, increasing the current global mean NUEr (~48%) is critical for food security and environmental protection. Optimized agricultural management—spanning nutrient (4R strategies: right source/type, rate, time, placement; enhanced efficiency fertilizers; organic–inorganic combinations), crop (rotations, cover crops, residue return), and soil practices (e.g., tillage)—can improve NUEr. However, NUEr and the effectiveness of practices vary widely among regions and income levels due to differences in fertilizer use intensity, regulations, and mechanization; for example, cropland NUEr averages 66–69% in the USA/EU but 21–35% in China/India, and can exceed 80% in sub-Saharan Africa due to low fertilizer use. Prior meta-analyses have often assessed single practices and have not fully accounted for how site conditions (climate and soil) mediate management impacts. This study aims to quantify how nutrient, crop, and soil management affect NUEr globally while explicitly accounting for site conditions (mean annual temperature and precipitation, soil organic carbon, clay content, and pH) using meta-analytical synthesis and meta-regression of primary field data.

The paper synthesizes evidence that high-income regions generally achieve higher NUEr due to better fertilizer management and regulatory frameworks, whereas middle-income countries often overapply low-cost urea leading to low NUEr; low-income regions may have high apparent NUEr due to underuse of N fertilizers. Prior meta-analyses have shown substantial benefits of enhanced efficiency fertilizers (e.g., ~20% increase in rice NUEr relative to urea) and have explored impacts of conservation practices and organic amendments, but typically in isolation. The literature indicates variability in responses across crop types (rice versus upland systems), climates, and soils, and emphasizes 4R nutrient stewardship. However, gaps remain in understanding interactive effects of multiple management practices with site conditions at global scale, and methodological differences among meta-analyses can bias aggregated effect estimates, motivating a unified meta-regression approach on primary data with explicit moderators.

Study design: The authors combined two complementary approaches: (1) a meta-model synthesizing 29 published meta-analyses on NUEr/N uptake responses to agronomic practices, and (2) a global meta-regression on primary field data compiled from 407 studies, yielding 2,436 paired observations for maize, wheat, and rice. Definition of outcome: NUEr (fertilizer N recovery efficiency) was defined using the N difference approach: (N uptake in fertilized plots − N uptake in unfertilized plots) / fertilizer N rate × 100%. Data collection: From the 29 meta-analyses, effect sizes and uncertainties for defined practices were extracted. Primary studies provided means, standard deviations or confidence intervals, sample sizes, locations, crop types, management practices, and NUEr in treatment and control. Where variances were missing, SDs were imputed from mean coefficients of variation. Site condition data (mean annual temperature, mean annual precipitation) were derived from CRU, and soil properties (soil organic carbon, clay content, pH) from SoilGrids, using study coordinates. Practices considered: Nutrient management (enhanced efficiency fertilizers, combined organic–mineral fertilizer, organic fertilizer, mineral fertilizer, right fertilizer placement, right rate, right timing); crop management (residue retention, cover cropping, crop rotation); soil management (zero tillage, reduced tillage). Effect sizes and models: For primary data, three effect size metrics were calculated: ROM (log ratio of means), MD (raw mean difference in absolute NUEr %), and SMD (standardized mean difference). Multivariate random-effects meta-regression (restricted maximum likelihood) with study as a random factor was used to account for non-independence and heterogeneity. Main effects and all two-way interactions between practices and site conditions (including crop type, N rate, SOC, clay, pH, MAT, MAP) were evaluated. ANOVA, AIC, and p-values assessed moderator contributions. For comparability, ROM and SMD results were converted to absolute NUEr changes. The MD model, which provided narrower confidence intervals and identified more interactions, was selected for global upscaling. Upscaling and mapping: Using the MD meta-regression, the authors predicted spatial variations in the potential average increase in absolute NUEr across global croplands under optimal combinations of practices (nutrient, crop, soil), reporting mean effects and 95% confidence bounds, and mapped contributions of each practice group.

- Management effects on NUEr (relative and absolute): - Meta-analytical synthesis indicated notable increases in relative NUEr for right fertilizer placement (~28%) and crop rotation (~38%). - Meta-regression on primary data found 7 of 12 practices increased relative NUEr, including enhanced efficiency fertilizers (EE, ~31%), combined fertilizer (~16%), right placement (~26%), right rate (~39%), right timing (~24%), residue retention (~24%), and cover cropping (~22%). Organic fertilizers and zero/reduced tillage generally decreased relative NUEr by ~1.2–9%. - In absolute terms (ROM-converted), NUEr increased up to ~10% for many practices: EE (+9.8%), combined fertilizer (+5.9%), right placement (+7.3%), right rate (+11%), right timing (+7%), residue retention (+8%), cover crops (+8%). Organic fertilizers, rotation, and zero/reduced tillage had limited or negative effects on absolute NUEr. - Role of site conditions and interactions: - Positive associations with higher SOC, soil pH, and MAP; negative associations with higher N application rate and clay content. MAT had a small negative effect. - The meta-regression explained 57–65% of variance in management-induced NUEr changes (MD: 65%; ROM: 57%; SMD: 63%). - MD parameter estimates indicated: selecting right fertilizer type or combined organic–mineral sources increased NUEr by +3.8 to +8.1%; optimizing placement, rate, and timing added ~+4.3%; residue retention +1.9%; rotation +1.6%; zero/reduced tillage −2.9%. Each kg N applied reduced NUEr by ~0.018% (marginal effect). Effects also varied by crop (wheat, maize, rice). - Global potential under optimal practices (MD model): - Combining nutrient practices (EE, combined, organic substitution, right placement, rate, timing) increased mean NUEr by ~27%. - Combining crop practices (residue retention, cover cropping, rotation) increased mean NUEr by ~6.6%. - Soil practices (zero/reduced tillage) changed mean NUEr by ~0.6% (near zero overall effect). - Overall, optimal combinations raised global absolute NUEr by ~30.3% on average (95% CI bounds: ~24.5% to ~36.6%), implying an increase from ~48% current mean to ~78%. - Regional patterns: Above-average gains in eastern Africa, central Asia, southern North America, central South America, and southern Australia; below-average gains in northern Europe, southern Asia, and eastern North America. Gains tended to be larger in middle-income regions (lower baseline NUEr) and smaller in many high-income regions (higher baseline NUEr).

The study demonstrates that optimizing nutrient management following 4R principles (right source, rate, time, placement) substantially improves NUEr by aligning N supply with crop demand, thereby reducing losses via volatilization, runoff, and leaching. Enhanced efficiency fertilizers slow N transformations to better match crop uptake windows, while partial substitution of mineral N with organic sources can improve soil structure, nutrient retention, and microbial activity, supporting sustained N availability; full substitution may depress NUEr, so combined use is preferable. Crop management practices such as residue retention and cover cropping improve soil organic matter and fertility, contributing more to NUEr than rotation alone in this synthesis, though rotations enhance resilience and multifunctionality. Zero or reduced tillage had small and often inconsistent effects on NUEr, potentially due to reduced aeration and seedling emergence, although such practices may confer other environmental benefits (biodiversity, soil structure) when combined with residue retention and rotations. The findings highlight strong modulation of management impacts by site conditions: higher SOC, pH, and precipitation tend to enhance NUEr responses, while high N rates, clay content, and higher temperatures can dampen them. Crop-specific physiology and water management (e.g., flooded rice vs upland systems) contribute to heterogeneity. Consequently, site-specific tailoring of practice combinations is essential to realize potential gains. Methodologically, the MD effect size provided more robust and interpretable estimates than ROM, supporting its use for upscaling. Overall, the results address the research question by quantifying both the average global potential and the spatial variability of management-driven NUEr improvements, demonstrating that optimized nutrient and crop practices can reconcile productivity with reduced N losses in many regions.

By integrating 29 meta-analyses and 2,436 paired observations from 407 field studies, the study shows that optimized agronomic management can raise global cropland NUEr by about 30% on average—from ~48% to nearly 78%. Nutrient management contributes the largest share (~27%), followed by crop management (~6.6%), while soil tillage changes have minimal direct effect on NUEr (~0.6%). The greatest improvements are expected where current NUEr is low or fertilizer is overapplied, while gains are smaller in regions with already high NUEr. Achieving these improvements will require coordinated, site-specific implementation of 4R nutrient strategies, judicious organic–mineral combinations, and supportive crop management, enabling productivity goals while curbing reactive N losses. Future research should refine site-specific recommendations, include broader crop systems and underrepresented regions, and integrate additional environmental outcomes and socioeconomic constraints into decision-support frameworks.

- Heterogeneity and methodological differences across included meta-analyses and primary studies (selection criteria, effect size definitions, statistical models) introduce uncertainty; the meta-model from published meta-analyses exhibited higher uncertainty than the primary-data meta-regression. - Site condition data were often not reported in primary studies and were inferred from gridded datasets (CRU, SoilGrids), which may not capture local variability. - The dataset focused on systems with sufficient fertilizer inputs; experiments involving nutrient mining (N uptake exceeding inputs) were not included, limiting generalizability to low-input contexts. - Study distribution was uneven geographically (majority in Asia, fewer in Africa and South America), and crop focus was on maize, wheat, and rice, potentially biasing global extrapolations. - Effect estimates for some practices (e.g., tillage) showed small and context-dependent impacts, and interactions among multiple simultaneous practices beyond two-way terms may be underrepresented.