Enhanced mitigation in nutrient surplus driven by multilateral crop trade patterns

The study addresses how multilateral crop trading patterns influence nitrogen (N) and phosphorus (P) surpluses and associated environmental risks. Rising global food demand (50–100% by 2050) implies increased fertilizer use, leading to groundwater contamination, eutrophication, and air pollution. While nutrient footprints and budgets have been widely studied, the response of nutrient surplus and surplus footprints to multilateral crop trade (MCT) remains unclear, particularly identifying trade patterns that mitigate rather than enhance surpluses. The purpose is to quantify spatiotemporal nutrient surplus footprints (NSF, PSF) and analyze their response to existing and optimal MCT scenarios, balancing environmental stress and economic returns. The study focuses on China and five Central Asian countries (Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan) due to their food security challenges, fertilizer management issues, environmental burdens, and growing but underdeveloped trade ties.

Prior research has examined agricultural N and P emissions, fate and transport, spatiotemporal patterns, and pollution control at global and regional scales. Nutrient footprint tools quantify lifecycle N and P associated with production and consumption. Nutrient surplus, reflecting inputs exceeding crop needs, has been emphasized due to its environmental risks, and improving nutrient use efficiency (NUE) is central to control. However, a gap exists connecting nutrient surplus metrics with multilateral trade patterns and virtual nutrient flows. The concept of nutrient surplus footprint (NSF/PSF) integrates surplus and footprint perspectives, enabling evaluation of environmental risk from remaining nutrients and benefits from stabilized nutrients, and linking with crop trade to analyze mitigation or enhancement effects due to virtual flows. Policymaking frameworks (e.g., SDGs) increasingly call for integrated nutrient management, but the role of MCT in surplus outcomes remains insufficiently quantified.

Scope and system boundary: The study quantifies national-scale N and P surpluses from cropland for China and five Central Asian countries over 1992–2018 using data for 144 crops aggregated to 12 categories. China’s surplus is first computed at provincial level and then aggregated to national level; Central Asian countries are assessed at national scale. Integrated Nutrient Surplus Footprint Evaluation Model (INSFEM): Calculates annual nutrient surplus and surplus footprints for cropland systems. Inputs include synthetic fertilizer (IN_fer), animal manure (IN_man), atmospheric deposition (IN_dep/IN_atm), and biological N fixation (IN_fix for N). Outputs are nutrient removal via harvested crops (N_har, P_har). Nutrient surplus is NS = IN_fer + IN_man + IN_dep − N_har; PS = IN_fer + IN_man + IN_dep − P_har. Surplus footprints are defined as NSF = NS / N_har and PSF = PS / P_har. Virtual nutrient flows are nutrients embodied in traded crops, shifting surplus implications across countries. Mitigation/enhancement due to trade: For trade of crop x from exporter j to importer i with traded mass T_{i,j,x}, mitigation metrics are NS_mitigation_{i,j,x} = T_{i,j,x} × (NSF_i − NSF_j) and PS_mitigation_{i,j,x} = T_{i,j,x} × (PSF_i − PSF_j). Negative values indicate mitigation (trade from low to high footprint regions), positive values indicate enhancement. Totals sum across all i, j, x. Trade-response analysis: Empirical regressions relate NSF/PSF to trade amount and trade structure for key crops (wheat, maize, fruits, vegetables, oil crops, etc.), testing significance (p < 0.05). Scenario design: Based on 2018 trade as baseline, four scenario groups are assessed: (1) Trade volume (TV): increase by 50%, 100%, 150% (TV1–TV3); decrease by 50%, 80%, 100% (TV4–TV6). (2) Trade structure (TS): adjust shares of fruits and vegetables by +50%, +100%, +150% (TS1–TS3) and −50%, −80%, −100% (TS4–TS6). These do not constrain demand or production capacity. (3) Optimal MCT single-objective (OP1): linear programming maximizing total system benefit with constraints on water/soil resources (irrigation water demand, planting area), demand and trade balance, water/soil benefit equalization, and nonnegativity. Decision variables include planting area, crop imports/exports. (4) Optimal MCT bi-level (OP2): upper-level objective minimizes inequality of water-land benefits (with constraints on inequality coefficient and virtual water efficiency); lower-level maximizes system benefit with constraints on planting area, irrigation water consumption, trading, and nonnegativity. Planning horizon 2020–2034, three 5-year periods. Data sources: Agricultural production/trade (FAOSTAT), fertilizer (IFA), meteorological and environmental data from Chinese Academy of Sciences data centers and National Tibetan Plateau Data Center. The regional surplus model can be implemented in Excel; optimization codes available on request.

• Spatiotemporal NSF/PSF: China’s average NSF (~2.0) and PSF (~2.8) exceed Central Asia averages (~1.2 and ~0.12) by ~67% and ~22x respectively. High footprints are concentrated in eastern/southern China and in Turkmenistan and Uzbekistan. Maize is a major contributor to NSF/PSF in China, Kyrgyzstan, and Uzbekistan; fruits, vegetables, cotton, and oil crops also drive high footprints in Turkmenistan and Uzbekistan. Some negative footprints (indicative of nutrient depletion) occur in Kazakhstan and Kyrgyzstan for certain crops (e.g., wheat, rice, roots/tubers). From 1992–2018, footprints increased across all six countries, notably in China, Turkmenistan, and Uzbekistan; crop contributions shifted from maize to fruits after 2011. Estimated fertilizer use efficiencies are low: ~33% (China) and ~45% (Central Asia) versus ~70% in developed countries.
• Trade relationships: Most countries were net importers except Kazakhstan; Tajikistan, Uzbekistan, and Kyrgyzstan had the largest net imports (avg. 346.3, 301.8, 154.0 kt, 1992–2018). China’s cotton imports peaked at 392.9 kt (2006) then declined to 82.6 kt (2018). Kazakhstan imported mainly vegetables and fruits, peaking at 418.9 kt and 485.3 kt (2013). Other countries’ imports were dominated by wheat (~84.5%).
• NSF/PSF response to trade amount/structure: Significant linear relationships (p < 0.05) were found between NSF/PSF and trade amounts/structures for key crops. In China, wheat/maize/fruits/oil crop trade amounts correlate strongly with NSF/PSF; in Central Asia, wheat/maize/fruits/vegetables relate to NSF, and wheat/pulses/roots-tubers/vegetables to PSF. Trade structure (shares of wheat/maize/oil crops in China and fruits/oil crops for NSF, maize/cotton/vegetables for PSF in Central Asia) also correlates.
• Virtual nutrient flows: Net virtual N and P flows intensified over time. Between Kazakhstan and China, net virtual N rose from 40 kt (1992–2000) to 225 kt (2011–2018); net virtual P rose from 40 to 413 kt (>10×), with P flows 38–84% higher than N in later periods.
• Historical trade effects: Past China–Central Asia trade generally enhanced NS and PS in China (e.g., peak total NS and PS enhancement in 2013: 2066 and 4575 kt; China–Kazakhstan: 1641 and 3753 kt in 2013). Some years showed mitigation with specific partners/crops, but overall patterns were not environmentally optimal.
• Scenario analysis (TV/TS): Increasing trade volume reduces Central Asia’s NSF/PSF (e.g., TV3: Central Asia NSF ~−2% from 1.38; PSF −5% from 0.30) and shifts China’s NS from enhancement (1170.5 kt in 2018) to mitigation (−530.3 kt under TV3); PS mitigation reaches −2131 kt. Adjusting trade structure (fruits/vegetables shares) produces small NSF/PSF changes but tends to increase NS enhancement and slightly reduce PS enhancement (e.g., TS6: NS enhancement rises to 1228.7 kt; PS enhancement declines to 1604 kt). Thus, trade volume increases are more beneficial for surplus mitigation than structural adjustments.
• Optimal MCTs (OP1/OP2): Both optimal models improve footprints and achieve stronger surplus mitigation than simple TV/TS changes. Under OP1, China’s NSF and PSF decline by ~3% (1.36→1.32) and ~2.4% (2.50→2.44); Central Asia’s NSF and PSF decline by ~16% (1.38→1.16) and ~37% (0.30→0.19). By period 3, NS transitions from +1170.5 kt to −705.8 kt; PS from +1741 kt to −2934 kt. Under OP2, mitigations are slightly smaller due to fairness constraints: NS to −571 kt; PS to −2809 kt. Both demonstrate that optimized multilateral trade can substantially mitigate nutrient surpluses.

The findings demonstrate that multilateral crop trade patterns significantly influence nutrient surpluses via virtual nutrient flows. Increasing trade volumes, particularly under optimized allocations that account for resource constraints and benefits, can redirect flows from low-footprint to high-footprint regions, achieving mitigation. In contrast, naive adjustments to trade structure (e.g., increasing fruits/vegetables shares without system optimization) can exacerbate N surplus while only modestly alleviating P surplus. Optimal MCTs (OP1/OP2) outperform simple volume/structure changes by jointly maximizing system-wide economic benefits and addressing equity in water–land resource benefits, leading to meaningful reductions in NSF/PSF and converting historical enhancement into mitigation for both N and P. These results address the research question by identifying mechanisms—trade volume scaling and optimal allocation—that drive mitigation and quantifying their impacts. The work highlights policy pathways: prioritize optimized, fairness-aware trade expansion; integrate nutrient performance (NSF/PSF) into trade and agricultural planning; and select crop portfolios with lower surplus footprints (e.g., wheat, roots/tubers) while improving management for high-footprint crops (maize, oil crops). Broader implications include embedding equity (e.g., Gini-based inequality) in trade frameworks, leveraging tools like GTAP for global extensions, and integrating NSF/PSF indicators into SDG monitoring to support sustainable nutrient management.

This study introduces INSFEM to quantify nutrient surplus footprints and links them to multilateral crop trade, revealing that historically, China–Central Asia trade generally enhanced nutrient surpluses but that optimized multilateral trade can reverse this effect. Scenario analyses show that increasing trade volume can shift China’s nitrogen and phosphorus balances from enhancement to mitigation, and optimal trade models (OP1/OP2) deliver the strongest improvements, cutting NSF/PSF and achieving large net mitigations (OP1: NS −705.8 kt, PS −2934 kt; OP2: NS −571 kt, PS −2809 kt). The main contributions are: (1) proposing an integrated framework (INSFEM) for surplus footprints and virtual nutrient flows; (2) empirically demonstrating trade–surplus relationships; (3) designing and evaluating optimal multilateral trade models that incorporate resource constraints and equity. Future research should expand spatial resolution, generalize models to global systems, integrate dynamic climatic and socioeconomic drivers, and develop optimization algorithms accommodating fairness and uncertainty for broader applicability.

• Data resolution: Central Asian analyses rely on national-scale data; finer-resolution inputs (e.g., field capacity, manure excretion, deposition) would improve precision but are difficult to obtain.
• System scope: The framework is demonstrated for six countries; generalization to global multilateral trade systems requires validation and increased modeling complexity.
• Model complexity and computation: Scaling to more countries and dynamic conditions demands improved model structures and solution algorithms, including fairness constraints.
• Exogenous drivers: Potential responses of nutrient surpluses to climate variability (e.g., precipitation changes) and evolving human activities (trade dynamics) are not explicitly modeled; their impacts and mitigation measures warrant future study.