Legume-Based Rotations Enhance Ecosystem Sustainability in the North China Plain: Trade-Offs Between Greenhouse Gas Mitigation, Soil Carbon Sequestration, and Economic Viability

Abstract

Reconciling agricultural productivity with greenhouse gas (GHG) mitigation remains a pivotal challenge for achieving climate-smart food systems. This study evaluates the capacity of legume-based crop rotations to balance economic viability, yield stability, and GHG reduction in the North China Plain. A two-year randomized complete block field experiment compared six cropping systems: conventional wheat–maize (WM) rotations and legume-integrated systems (wheat–soybean, WS; wheat–soybean–maize, WSM), under fertilized and unfertilized regimes. Results revealed that nitrogen fertilization increased cumulative N₂O emissions and global warming potential (GWP), with seasonal peaks occurring post-fertilization. Legume systems enhanced CH₄ uptake but showed no significant effect on N₂O emissions compared to conventional systems. N₂O fluxes correlated positively with soil moisture and soil temperature, while CH₄ uptake increased with soil moisture alone. Soybean phases reduced short-term yields by 32–52% relative to the maize yield of conventional systems, but boosted subsequent wheat/maize productivity by 2–47% through hydraulic redistribution and N priming. The wheat–soybean rotation with 200 kg N ha⁻¹ (WS200) achieved optimal sustainability, delivering the highest net profit (8061.56 USD ha⁻¹) alongside a 9% reduction in global warming potential (3980.21 kg CO_2-eq ha⁻¹) versus conventional systems. These findings provide actionable insights for sustainable intensification in global cereal systems, demonstrating that strategic legume integration can advance both food security and climate goals.

1. Introduction

The unabated rise in atmospheric concentrations of carbon dioxide (CO₂) and potent non-CO₂ greenhouse gases (GHGs), particularly nitrous oxide (N₂O) and methane (CH₄), poses a critical threat to global climate stability [1]. Given that N₂O and CH₄ possess 298 and 34 times the global warming potential of CO₂ over a century, respectively [2], mitigating anthropogenic emissions has become a planetary imperative. According to FAO data from 2023, agriculture contributes approximately 16.2% of global anthropogenic GHG emissions. In intensive farming regions such as the North China Plain (NCP), synthetic fertilizer application and intensive tillage are responsible for over 80% of agricultural N₂O emissions [3,4]. Consequently, reducing agricultural GHG emissions is therefore critical for climate resilience. However, a fundamental challenge lies in achieving emission reductions without compromising crop productivity, particularly in vital breadbasket regions essential for global food security—a central tenet of sustainable intensification [5,6].

Current research on agricultural GHG mitigation has predominantly focused on incremental modifications to nitrogen (N) management within conventional monoculture systems [7,8]. These approaches, however, frequently lead to significant trade-offs and inconsistencies. For example, short-term straw incorporation has been shown to increase N₂O emissions by 35–60% in wheat–maize systems [9], and conventional N fertilization can amplify N₂O fluxes by 176–222% compared to unfertilized soils [10]. Similarly, in double-rice systems, even reduced N application can increase seasonal N₂O emissions by 42%, while full N rates may increase CH₄ emissions by 26% [11]. These contradictions highlight a critical research gap: a lack of integrated strategies that effectively address GHG emissions, crop productivity, and economic viability simultaneously.

Legume–cereal rotations have emerged as a promising strategy to decouple agricultural productivity from synthetic N inputs by utilizing biological nitrogen fixation (BNF). These systems can potentially reduce N₂O emissions by 35–80% while maintaining or even improving yields [12,13,14]. By substituting chemical fertilizers with ecological processes, legume integrations align with principles of cleaner production, thereby curbing resource depletion and pollution [15]. Despite this potential, the scalability of legume-based systems is hindered by unresolved scientific questions. First, legumes may paradoxically stimulate N₂O emissions via rhizodeposition of labile carbon and nitrogen [16], yet the magnitude of this effect under field conditions remains poorly quantified. Second, the impacts of legume rotations on CH₄ fluxes—particularly in non-flooded upland systems like those in the NCP—are inconsistently reported, with some meta-analyses indicating negligible crop-type effects [17]. Most critically, comprehensive comparisons between legume-integrated systems and conventional rotations, especially under equivalent N fertilization regimes, are scarce. This lack of direct comparison limits a mechanistic understanding of GHG emission drivers and impedes robust cost–benefit analyses crucial for farmer adoption [18].

To bridge these knowledge gaps, this study was designed to address two key questions: (1) How do legume-based rotations modulate the net GHG budget, particularly when accounting for the trade-offs between potential N₂O reduction and CH₄ stimulation? (2) Can these systems achieve dual economic and ecological advantages by reducing input costs while maintaining yields in intensive cropping landscapes? Through a multi-year field experiment and life-cycle assessment, we evaluate the potential of legume-cereal rotations in the NCP to reduce net GHG emissions and increase farm net income. Our findings aim to inform the design of cropping systems that reconcile climate resilience with food security, a cornerstone of sustainable agriculture.

2. Materials and Methods

2.1. Experimental Site

The study was conducted over a two-year period (2021–2023) in Shanhoujie Village, Fengyang County, Anhui province, China (32°53′ N, 117°33′ E), a region characterized by a north-subtropical humid monsoon climate. The mean annual temperature is 15.7 °C, and the mean annual precipitation is 900 mm, with approximately 70% of the rainfall occurring during the summer monsoon season (June–August). The soil at the experimental site is classified as Yellow-Brown Earth (FAO classification), with a texture composition of 16.5% clay, 28.8% silt, and 54.7% sand. Prior to the experiment, initial soil properties (0–20 cm depth) were as follows: organic matter (13.6 g kg⁻¹), total nitrogen (1.68 g kg⁻¹), available phosphorus (23.95 mg kg⁻¹), available potassium (91.63 mg kg⁻¹), and pH 6.6 (soil:deionized water ratio = 1:2.5, w/v).

Agronomic practices at the field site were consistent with local conventional farming systems. The tillage regime was conventional tillage, involving annual ploughing to a depth of approximately 20 cm before the sowing of each crop. Crop stubble and other residual organic materials were uniformly incorporated into the soil during this ploughing operation. The primary crops in the regional cropping system are winter wheat and summer maize. Given that the precipitation distribution is uneven and often insufficient to meet crop water demands, especially during critical growth stages of winter wheat, irrigation is essential. The irrigation method was flood irrigation, with two scheduled events: the first at the jointing stage and the second at the flowering stage. Each irrigation application delivered approximately 60 mm of water, totaling 120 mm per wheat season. This practice aligns with the optimized water management strategies recommended for the wheat–maize system in the NCP to maintain yields while improving water use efficiency.

2.2. Experimental Design

A randomized complete block design was implemented with six treatments and three replicates (18 plots total; plot area: 16.8 m², 4.2 m × 4 m), and a 1.5 m buffer distance between plots to avoid cross-treatment interference. Treatments included (

Table 1. The entire experimental procedure. GS1, GS2, GS3 and GS4 represent the first, second, third, and fourth growing season, respectively.

Season	Sequence	Crop	Planting Date	N Fertilization			Harvest Date
				Date	Type	kg N ha−1
GS1	WM0	maize	13 June				25 September
	WM330	maize		11 June	Basal	120
				15 August	Top dressing	60
	WS0	soybean
	WS200	soybean		11 June	Basal	50
	WSM0	soybean
	WSM265	soybean		11 June	Basal	50
GS2	WM0	wheat	19 October				26 May
	WM330	wheat		17 October	Basal	100
				25 March	Top dressing	50
	WS0	wheat
	WS200	wheat		17 October	Basal	100
				25 March	Top dressing	50
	WSM0	wheat
	WSM265	wheat		17 October	Basal	100
				25March	Top dressing	50
GS3	WM0	maize	15 June				20 September
	WM330	maize		13 June	Basal	120
				9 August	Top dressing	60
	WS0	soybean
	WS200	soybean		13 June	Basal	50
	WSM0	maize
	WSM265	maize		13 June	Basal	120
				9 August	Top dressing	60
GS4	WM0	wheat	22 October				27 May
	WM330	wheat		20 October	Basal	100
				19 March	Top dressing	50
	WS0	wheat
	WS200	wheat		20 October	Basal	100
				19 March	Top dressing	50
	WSM0	wheat
	WSM265	wheat		20 October	Basal	100
				19 March	Top dressing	50

):

(1) WM0: Wheat–maize rotation, unfertilized (0 kg N ha⁻¹ yr⁻¹).

(2) WM330: Wheat–maize rotation, conventionally fertilized (maize: 120 kg N ha⁻¹ as basal 15:15:15 NPK + 60 kg N ha⁻¹ urea top dressing; wheat: 100 kg N ha⁻¹ basal 15:15:15 NPK + 50 kg N ha⁻¹ urea).

(3) WS0: Wheat–soybean rotation, unfertilized (0 kg N ha⁻¹ yr⁻¹).

(4) WS200: Wheat–soybean rotation, fertilized (soybean: 50 kg N ha⁻¹ basal 15:15:15 NPK; wheat: 100 kg N ha⁻¹ basal 15:15:15 NPK + 50 kg N ha⁻¹ urea).

(5) WSM0: Wheat–soybean–maize rotation, unfertilized (0 kg N ha⁻¹ yr⁻¹).

(6) WSM265: Wheat–soybean–maize rotation, fertilized (soybean: 50 kg N ha⁻¹; wheat: 150 kg N ha⁻¹; maize: 180 kg N ha⁻¹; all as 15:15:15 NPK and urea split applications).

In alignment with local agricultural practices, NPK compound fertilizer (N:P₂O₅:K₂O = 15%:15%:15%) was applied as basal fertilizer to supply balanced nutrients during the early crop growth stage. Urea (46% N) was used as top dressing to meet nitrogen demands during the rapid growth phase. The crop cultivars employed in the study were winter wheat (Triticum aestivum L. cv. Yannong 19), maize (Zea mays L. cv. Zhengdan 958), and soybean (Glycine max L. cv. Zhonghuang 37). Standard agronomic practices (planting density: maize 67,500 plants ha⁻¹, wheat 2.25 × 10⁶ seeds ha⁻¹, soybean 195,000 plants ha⁻¹) were implemented consistently across all treatments. Uniform irrigation, pest control, and other field management operations were applied throughout the growing season. Weeds were controlled mechanically via inter-row hoeing at the seedling stage and supplemented with glyphosate (1.5 kg ha⁻¹) applied pre-emergence where necessary.

2.3. Gas Sampling and Analysis

N₂O and CH₄ fluxes were measured using static closed chambers (50 cm × 50 cm × 50 cm, PVC) equipped with water-sealed bases, thermal insulation (sponge and aluminum foil), and an internal fan to homogenize air. Gas samples (60 mL) were collected weekly between 08:00 and 11:00 a.m., with increased frequency post-fertilization and rainfall, at 0-, 8-, and 16-min intervals. Three static chambers were deployed per plot, positioned randomly but avoiding plot edges (>0.5 m from boundaries).

Flux rates (μg or mg m⁻² h⁻¹) were calculated as follows:

$$\mathrm{F}=\mathrm{\rho }\times {\displaystyle \frac{\mathrm{V}}{\mathrm{A}}}\times {\displaystyle \frac{\mathrm{d}\mathrm{c}}{\mathrm{d}\mathrm{t}}}\times {\displaystyle \frac{273}{273+\mathrm{T}}}$$

where ρ is the gas density (CH₄: 0.54 g L⁻¹; N₂O: 1.25 g L⁻¹), V is the chamber volume (0.125 m³), A is the soil surface area (0.25 m²), dc/dt is the gas concentration gradient, and T is the chamber temperature (°C).

Cumulative emissions (E, kg ha⁻¹ yr⁻¹) were derived by linear interpolation between sampling intervals:

$$\mathrm{E}=\sum [{\displaystyle \frac{\left({\mathrm{F}}_{\mathrm{i}+1}+{\mathrm{F}}_{\mathrm{i}}\right)}{2}}\times ({\mathrm{T}}_{\mathrm{i}+1}-{\mathrm{T}}_{\mathrm{i}})]$$

where F_i is the gas emission rate, i is the sampling number, and T is the sampling time.

Global warming potential (GWP, kg CO_2-eq ha⁻¹ yr⁻¹) was calculated as follows [19]:

GWP = 298 × N₂O + 34 × CH₄

where the greenhouse effects of N₂O and CH₄ are 298 and 34 times that of CO₂, respectively.

2.4. Soil Characteristics

Soil samples were collected from 0 to 20 cm depth (plough layer) using a zigzag sampling strategy across each plot. Soil nitrate (NO₃⁻-N) and ammonium (NH₄⁺-N) were quantified via flow injection analysis (SmartChem 200, AMS Alliance, France). Total nitrogen (TN) and dissolved organic carbon (DOC) were determined using a CN elemental analyzer (Vario Macro Cube, Elementar, Germany). Available P was measured using the Olsen bicarbonate extraction method. Soil temperature and moisture were recorded during gas sampling. Soil pH was determined by pH meter with a ratio of 1:2.5 (air dried-soil: deionized water solution).

2.5. Crop Yield, Biomass and Economic Benefit Analysis

After harvesting plants, the seeds were stripped and dried, and the remaining parts were oven-dried at 65 °C for 48 h to reach constant weight. Then, the biomass of the plants, grain yield and economic benefits under each crop rotation model were obtained by weighing. Meanwhile, the economic benefit was calculated as follows (1 USD = 7.0 RMB):

Economic benefit = (grain yield × crop price) − (fertilizer input × fertilizer price)

2.6. Statistical Analysis

Data were analyzed using IBM SPSS Statistics (V27) and visualized with Origin software (V2022). Treatment effects were assessed via one-way ANOVA (Analysis of Variance) with Tukey’s HSD post hoc tests (α = 0.05). Pearson correlations linked gas fluxes to soil parameters. Results are reported as mean ± standard error (SE; n = 3).

3. Results

3.1. Soil Biogeochemistry Properties and Nitrogen Dynamics

Soil moisture and temperature exhibited clear seasonal patterns, consistent across all treatments (Figure 1). Soil moisture peaked during the post-monsoon period (August–September) at 42–48% (v/v) and declined to minima of 18–22% (v/v) in winter. Soil temperature ranged from 2.0 °C in January to 30.8 °C in July, reflecting typical annual cycles for the region.

Legume-based rotations significantly influenced soil nitrogen availability and fertility after a two-year experiment (

Table 2. Soil properties under different treatments at the end of the growing season. Different letters mean differ significantly (p < 0.05).

	pH	NH4+-N mg kg−1	NO3−-N mg kg−1	DOC g kg−1	TN g kg−1
WM0	6.63 ± 0.09 a	14.03 ± 1.07 b	51.24 ± 1.89 b	13.43 ± 0.12 b	1.58 ± 0.04 c
WM330	6.33 ± 0.03 b	18.73 ± 0.60 a	60.16 ± 1.74 a	14.27 ± 0.20 a	1.87 ± 0.06 a
WS0	6.57 ± 0.07 ab	15.25 ± 1.08 ab	50.38 ± 1.87 b	13.50 ± 0.17 b	1.73 ± 0.03 ab
WS200	6.40 ± 0.06 ab	16.97 ± 0.74 ab	56.37 ± 1.98 ab	13.93 ± 0.09 ab	1.81 ± 0.01 ab
WSM0	6.53 ± 0.03 ab	14.50 ± 0.41 b	52.35 ± 1.71 ab	13.63 ± 0.09 ab	1.68 ± 0.02 bc
WSM265	6.40 ± 0.06 ab	17.57 ± 0.57 ab	57.76 ± 1.27 ab	14.07 ± 0.15 ab	1.84 ± 0.01 a

). Compared to the unfertilized control (WM0), the conventionally fertilized wheat–maize system (WM330) resulted in substantial increases in NH₄⁺-N (+33.50%), NO₃⁻-N (+17.41%), DOC (+6.25%), and TN (+18.35%) (p < 0.05). This enhancement in soil N pools within legume-integrated systems is attributable to biological nitrogen fixation, which effectively offsets the demand for synthetic fertilizers.

3.2. N₂O and CH₄ Flux Dynamics and Driving Factors

Despite no significant differences in instantaneous N₂O flux magnitudes between treatments (Figure 2a), pronounced temporal dynamics were observed. Fluxes increased sharply by 5- to 10-fold (reaching 11.68–595.49 μg m⁻² h⁻¹) within 7 days following basal nitrogen application, followed by a gradual decline influenced by rainfall and tillage events. Cumulative annual N₂O emissions were significantly lower in the unfertilized wheat–maize system (WM0, 323.6 mg m⁻²) compared to fertilized treatments (Figure 3a). Unfertilized legume rotations (WS0 and WSM0) increased emissions by 7.91% and 28.1%, respectively, relative to WM0. Fertilization markedly elevated N₂O emissions, with conventional wheat–maize (WM330) and wheat–soybean (WS200) systems producing 2.3-fold and 1.7-fold more N₂O than their unfertilized counterparts (p < 0.001). Seasonal analysis showed that summer (June–August) emissions exceeded winter fluxes by a factor of 2.3 to 4.1 (p < 0.01), correlating with higher soil temperatures (25–30 °C) and microbial activity.

In contrast, CH₄ fluxes demonstrated clear treatment-specific patterns (Figure 2b). All cropping systems acted as net CH₄ sinks, with uptake rates ranging from −0.49 mg m⁻² h⁻¹ (WM0) to −0.18 mg m⁻² h⁻¹ (WSM265). Legume-integrated systems significantly enhanced CH₄ uptake; specifically, the wheat–soybean maize rotation (WSM265) achieved 63% greater uptake than the conventional WM0 system (p < 0.05), likely due to improved soil aeration from legume root activity (Figure 3b). Cumulative CH₄ uptake was strongest in winter (December–February), exceeding summer uptake by 58–72% (p < 0.01), which corresponded with lower soil moisture levels (15–25% in winter vs. 35–45% in summer).

Correlation analyses revealed that N₂O fluxes were weakly but significantly positively correlated with both soil moisture (r² = 0.18, p < 0.001; Figure 4a) and soil temperature (r² = 0.17, p < 0.001; Figure 4b). CH₄ uptake showed a stronger negative correlation with soil moisture (r² = 0.24, p < 0.001; Figure 4c) but no significant relationship with temperature (Figure 4d). Notably, legume rotations strengthened the positive correlation between N₂O emissions and soil total nitrogen content (Figure 5), suggesting that N mineralization stimulated by rhizodeposition is a key driver of emissions.

3.3. Emission Factors and Global Warming Potential

Seasonal emission factors (EFs) for legume rotations (WS, WSM) were higher than those of conventional systems (0.33–2.06% vs. 0.65–1.01% of applied N; p < 0.05; Figure 6a). However, on an annual basis, EFs were not significantly different across systems (0.66–0.95%, p > 0.05), indicating compensatory mechanisms occurring over different crop growth phases. Fertilization regime had no significant impact on EFs (p > 0.1), underscoring the dominant role of inherent soil N legacy effects over short-term N inputs.

Synthetic N fertilization increased GWP by 67–148% compared to unfertilized controls (Figure 6b). The conventional wheat–maize system (WM330) had the highest GWP (4524.88 kg CO_2-eq ha⁻¹ yr⁻¹), followed by the wheat–soybean–maize rotation (WSM265; 4110.50 kg CO_2-eq ha⁻¹ yr⁻¹). The integration of legumes significantly modulated the GWP; the wheat–soybean system (WS200) reduced N₂O-dominated radiative forcing by 18–22% per unit yield, demonstrating a lower emission intensity (0.28–0.32 kg CO_2-eq ha⁻¹ grain) compared to monoculture systems (0.41–0.49 kg CO_2-eq kg⁻¹ grain).

3.4. Crop Yield, Economic Performance, and Integrated Sustainability

Fertilized systems consistently outperformed unfertilized controls in terms of crop yield (Figure 7a). Maize in the WM330 rotation achieved the highest productivity (9038.04 ± 272.04 kg ha⁻¹), while the wheat–soybean system (WS200) yielded the highest wheat production (7542.54 ± 165.01 kg ha⁻¹). Although the soybean phase initially reduced short-term output by 35–52% relative to maize in conventional systems, it significantly enhanced the yield of subsequent wheat and maize crops by 11–47% (p < 0.05), demonstrating a clear rotational synergy.

Economic returns closely followed yield trends (Figure 7b). Fertilized systems generated 1.8- to 2.0-fold higher net profits than unfertilized controls. The wheat–soybean rotation (WS200) emerged as the most sustainable option, delivering the highest net profit (8061.56 USD ha⁻¹) while reducing GWP by 12% relative to the conventional WM330 system (Figure 8). This legume-based system successfully balanced high annual grain yield (12.12–14.39 t ha⁻¹), low emission intensity (0.28–0.32 kg CO_2-eq kg⁻¹ grain), and improved input efficiency, establishing it as a climate-resilient paradigm for the NCP.

4. Discussion

4.1. Legume Rotations Modulate GHG Emissions Through Competing Pathways

Our results demonstrate that legume-based rotations (WS, WSM) exert divergent effects on N₂O and CH₄ fluxes, challenging the presumption of their universal mitigation efficacy. Contrary to the previous study reporting 39% N₂O reduction in legume systems [15], we observed no significant reduction in cumulative N₂O emissions between legume-integrated and conventional rotations (Figure 3a). Three non-exclusive mechanisms may explain this discrepancy: (1) Climatic Constraints on Nitrogen Cycling: The semi-arid conditions of our study site (900 mm annual precipitation) likely limited microbial nitrification and denitrification, despite additional N inputs from soybean residues. This observation aligns with findings by Barton et al. [20], who reported an exponential decrease in N₂O emissions from legume systems when annual rainfall fell below 1200 mm. (2) Soil Biochemical Inertia: Legume rotations failed to alter key N₂O drivers—DOC, pH, and NH₄⁺-N—within the topsoil (Figure 5). As soybean roots primarily occupy shallow horizons [21], their rhizodeposits may insufficiently perturb subsoil N pools critical for denitrification. (3) Temporal Compensation: Elevated emission factors during legume phases (1.60–2.06% vs. 0.86% in WM; Figure 6a) were offset by post-legume emission declines, resulting in net annual parity.

Legume rotations significantly attenuated CH₄ uptake compared to conventional systems (Figure 3b), contradicting reports of enhanced methanotrophy in legume-enriched soils [22]. We propose two potential mechanisms: (1) Moisture-Mediated Methanotroph Inhibition: Strong linear correlations between CH₄ uptake and soil moisture (r² = 0.24; Figure 4c) suggest that legume-induced water retention (via reduced evapotranspiration) suppressed methanotroph activity. This aligns with Gutenberg et al. [23], who documented a 457 μg m⁻² h⁻¹ CH₄ flux increase per 10% soil moisture rise. (2) Microbial Community Shifts: Though unmeasured here, legume root exudates (e.g., flavonoids, organic acids) may have reshaped methanotroph communities, as observed in analogous systems [24]. Future studies should couple gas flux measurements with pmoA gene quantification to test this hypothesis.

Despite the complex interplay between N₂O and CH₄ fluxes, legume systems reduced net GWP by 9–12% compared to conventional rotations (Figure 6b). This net benefit primarily stems from avoiding emissions associated with the production and application of synthetic N fertilizers, corroborating life-cycle assessments of legume-based agroecosystems [25,26].

4.2. Yield Synergies and Soil Legacy Effects

While soybean phases reduced short-term yields by 35–52%, subsequent wheat and maize productivity increased by 11–47% (Figure 7a), confirming the rotational synergy observed in global syntheses [27,28]. This legacy effect appears to be driven by two mechanisms: (1) Hydrological Buffering: Soybean’s shallow root architecture preserved subsoil moisture for deeper-rooted subsequent crops (wheat: ~1.2 m; maize: ~2.0 m), mitigating drought stress during grain filling—a critical advantage under climate variability [29]. (2) Nitrogen Priming: Symbiotic N fixation (BNF) by soybeans contributed 45–68 kg N ha⁻¹ (estimated via N difference method), reducing synthetic N demand for subsequent wheat by 20–25%. This aligns with isotopic tracer studies showing 25–40% of legume-fixed N transfers to subsequent cereals [30].

The observed yield penalty in legume phases mirrors Sánchez-Navarro et al. [18], who attributed it to lower photosynthetic efficiency in legumes versus C4 cereals. However, long-term soil health benefits—improved aggregate stability (+18%), microbial biomass C (+32%), and cation exchange capacity (+15%) in legume systems—suggest progressive yield convergence over multi-decadal scales [31].

4.3. Reconciling Productivity and Sustainability via Legume Integration

Long-term traditional crop rotations, such as wheat–maize systems, have been shown to degrade soil quality and exacerbate GHG emissions [32]. While synthetic N fertilizers have nearly doubled global grain yields, they have also increased reactive N losses, contributing to environmental pollution [33]. Overapplication of N fertilizers in wheat–maize rotations is a widespread issue, often failing to enhance yields while causing soil acidification and nutrient imbalances [34]. In contrast, legume-based rotations offer a sustainable alternative, mitigating these challenges through biological nitrogen fixation (BNF) and improved nutrient cycling [30,31].

In terms of ecological benefits, legumes significantly reduce the demand for external N fertilizers compared to cereals [18], thereby reducing the overall N inputs of the rotation system. This reduction in synthetic fertilizer use enhances N use efficiency, as demonstrated in barley–soybean rotations [35,36]. Furthermore, decreased N application directly reduces N₂O emissions, a major contributor to agricultural GHG emissions [37]. Additionally, legumes also improve soil fertility through BNF, which fixes atmospheric N₂ and addresses soil nutrient imbalances, supporting integrated soil fertility management (ISFM) [38]. For instance, approximately 33% of total legume N (ranging from 13% to 74%) remains in the soil, enriching SOC and addressing the “soil C dilemma” [16,39].

In terms of economic benefits, the reduced dependency on synthetic fertilizers lowers input costs, improving the net profitability of legume rotations compared to traditional systems. Additionally, legume residues, characterized by high nitrogen content and low carbon-to-nitrogen ratios, decompose rapidly, releasing nutrients for subsequent crops [40]. This enhances nutrient availability and supports higher yields in follow-on crops, further boosting economic returns. Generally, legume-based rotations provide a dual benefit of ecological sustainability and economic viability, making them a promising strategy for sustainable agriculture.

5. Conclusions

This study provides a comprehensive, system-level quantification of the agroecological trade-offs and synergies associated with legume-integrated rotations in the intensive cereal systems of the NCP. Our findings make several key contributions to the literature on sustainable agricultural intensification.

First, we offer a mechanistic explanation for the nuanced effect of legumes on greenhouse gas budgets. While legume rotations did not significantly alter cumulative N₂O emissions compared to conventional wheat–maize systems, they induced a distinct temporal shift: elevated N₂O emission factors during legume phases (1.60–2.06% vs. 0.86% in WM) were offset by post-legume emission declines, reflecting delayed N mineralization from root residues, which clarifies contradictory reports in the literature and underscores the importance of assessing emissions at the multi-seasonal, system-level rather than based on single crops or seasons. Furthermore, CH₄ uptake decreased by 18–63% in legume systems—a paradoxical outcome, likely driven by legume-induced soil moisture retention, which highlights a critical trade-off that must be managed. Despite this, net GWP reductions of 22–31% were achieved primarily through avoided synthetic N fertilizer emissions, demonstrating a clear pathway for decarbonizing intensive agriculture. Second, our research elucidates the agroecological mechanisms underpinning rotational synergy. While legume rotations slightly reduced overall yields compared to conventional systems, they significantly enhanced subsequent wheat and corn productivity. This yield boost is attributed to improved soil water conservation, nutrient cycling, and biological nitrogen fixation. This provides a solid, evidence-based rationale for using legumes as a tool for ecological intensification, where yield stability is achieved through the enhancement of ecosystem services rather than purely through external inputs. Third, we identify the wheat–soybean rotation (WS200) as an optimal system that successfully reconciles economic and ecological objectives. This system delivered the highest net profit alongside a reduced GWP, proving that environmental sustainability and economic viability are not mutually exclusive goals. By reducing synthetic fertilizer dependency, lowering production costs, and improving soil health, this model offers a scalable, climate-resilient paradigm for the NCP and similar global breadbaskets.

In conclusion, these findings underscore the potential of legume-based rotations to address the dual challenges of climate change and food security. By reducing reliance on synthetic fertilizers, enhancing soil fertility, and maintaining productivity, legume rotations offer a scalable solution for sustainable agriculture in similar agroecological regions. Future research should focus on optimizing legume–cereal ratios and nitrogen management strategies across diverse climatic gradients to further enhance the resilience and adaptability of these systems.