The field experiment was conducted at the Water-Saving Irrigation Experimental Station, Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas (Ministry of Education), Northwest A&F University. The experimental site is located in the Yangling Demonstration Zone, Xianyang, Shaanxi Province, Northwest China (108°24’ E, 34°20’ N; Figure 1). This area is situated in the western part of the Guanzhong Plain with a long-term average annual temperature of 12.9 °C. The long-term mean annual precipitation at the study site is 632 mm (1981–2020 climatology), whereas the experimental year (2023–2024) received 578.7 mm of precipitation according to observations from the on-site meteorological station. To ensure realistic soil-moisture scenarios, the three irrigation levels (W0, W1, W2) were designed after analyzing the 1981–2020 rainfall distribution: about 70% of annual precipitation falls between June and September, often as high-intensity events ≥ 10 mm day^-¹. The irrigation schedules therefore mimicked dry-year (W0), average-year (W1) and wet-year (W2) moisture regimes by supplementing 0, 100 and 300 mm, respectively, during the spring and early-summer gaps when alfalfa root uptake is highest. The timing and quantity of irrigation were therefore aligned with the seasonal rain. Precipitation and temperature data for the study area in the experimental period were obtained from a meteorological station within the experimental site (Figure 2). The soil type is classified as Haplic Cambisol characterized by a clay-loam texture and a field capacity of ~22.4% (gravimetric moisture content) in the 0–20 cm surface layer. The basic physicochemical properties of the surface soil were measured prior to sowing: pH (potentiometric method), 8.5; total nitrogen (Kjeldahl method), 1.0 g kg^-¹; available phosphorus (Olsen method), 15 mg kg^-¹; nitrate nitrogen (continuous flow analysis), 23 mg kg^-¹; ammonium nitrogen (continuous flow analysis), 8 mg kg^-¹; total organic matter (dichromate oxidation method, Walkley-Black), 15.5 g kg^-¹; and bulk density (cutting ring method), 1.48 g cm^-³.

The experiment was carried out between October 2023 and October 2024. To ensure that the experimental design aligns with local agronomic practices, the irrigation and fertilization levels were established based on regional water availability and the actual water and nitrogen requirements of alfalfa in the study area. The Yangling region has an average annual precipitation of approximately 630 mm, mostly occurring from June to September. The irrigation levels (W0: 0 mm, W1: 100 mm, W2: 300 mm) were all within the conventional supplemental irrigation range for local alfalfa production (100–350 mm). The irrigation treatments were therefore designed primarily for comparison purposes under different water conditions in relation to greenhouse gas (GHG) emissions and yield performance.

The nitrogen application rates (N0: 0 kg ha^-¹, N1: 60 kg ha^-¹, N2: 120 kg ha^-¹, N3: 180 kg ha^-¹) were based on the locally recommended fertilization range for alfalfa (60–180 kg ha^-¹), covering low to high levels typically observed in production. Such a design provides regionally applicable insights into the balance between agronomic productivity and environmental sustainability under variable water–fertilizer regimes. A total of 12 water-nitrogen management schemes were established (Table 1), with each replicated three times across 36 plots. The plots (2 m × 2 m each) were separated by 1 m-wide isolation strips and furrows. Seeds of the alfalfa variety WL63HQ (Autumn Sleep Grade 5) were obtained from Zhengdao Seed Industry Co., Ltd. (Beijing, China) and sown at a rate of 25 kg ha^-¹ with a row spacing of 20 cm. The fertilizers used in the experiment included superphosphate (46% P₂O₅, 75 kg ha^-¹), potassium fertilizer as K₂O (90 kg ha^-¹), and urea (46% N). Urea (46% N) were applied with a basal-to-topdressing ratio of 4:6. Basal nitrogen fertilizer (40%) was applied at sowing on October 15, 2023, followed by two topdressings (30% each) on March 14 (first harvest) and June 4 (second harvest), 2024. Irrigation took place on March 14 and April 14 (first harvest), June 4 (second harvest), and August 11 (third harvest), 2024. Alfalfa crops were mowed at the initial flowering stage, with three harvests scheduled annually on May 15, July 20, and September 27, 2024. Standard agronomic practices were maintained consistently across all treatments to isolate the effects of water and nitrogen. Prior to sowing, rotary tillage was performed to prepare a uniform seedbed. Basal fertilizers (Phosphate and Potassium) were incorporated into the soil during tillage to ensure adequate nutrient supply. During the growing season, integrated pest and weed management (using standard insecticides and manual weeding) was applied uniformly when necessary to prevent biotic stress. Harvest operations were conducted mechanically at a stubble height of 3 cm.

The N₂O emission factor (EF) was estimated using Equation 7 (IPCC, 2014):

where EF represents the N₂O emission factor (%); GF_F and GF_N are the cumulative N₂O emissions from fertilized and unfertilized treatments, respectively (kg N ha^-¹); and N_F is the nitrogen rate (kg N ha^-¹).

To comprehensively evaluate the environmental impact of the alfalfa production system, the net global warming potential (GWP, kg CO₂-eq ha^-¹) was calculated using a Life Cycle Assessment (LCA) approach. This method integrates both direct soil GHG emissions and indirect emissions associated with agricultural inputs (fertilizer production, irrigation electricity, and diesel fuel). The calculation is expressed as shown in Equations 8 and 9:

where GWPinputs represents the GWP from agricultural inputs (kg CO₂-eq ha^-¹). The carbon emission factors ( EFi) were adopted from recent literature on Chinese agriculture: 8.3 kg CO₂-eq kg^-1 for N fertilizer production, 0.79 kg CO₂-eq kg^-1 for P₂O₅, 0.55 kg CO₂-eq kg^-1, for K₂O, 2.66 kg CO₂-eq mm^-1 for irrigation water (electricity), and 55.4 kg CO₂-eq ha^-¹ for diesel fuel consumption during tillage and harvest (Chen et al., 2011; Zhang et al., 2017b).

Regarding the soil component ( GWPsoil), the calculation included cumulative CH4 and N2O emissions converted to CO2-equivalents over a 100-year horizon (with warming potentials of 28 and 265, respectively), as shown in Equation 10:

where GWPsoil represents the GWP from direct soil fluxes (kg CO₂-eq ha^-¹). It is important to note that chamber-measured soil CO₂ fluxes were excluded from the calculation. While soil respiration is a significant carbon flux, the static dark chamber method captures only ecosystem respiration (an output) without accounting for photosynthetic carbon uptake (an input). Including these fluxes would erroneously inflate the net carbon balance (Shang et al., 2019). Thus, our approach represents a conservative estimate focusing on non-CO₂ gases and input-based emissions.

Greenhouse gas emission intensity (GHGI) was defined as the LCA-based GWP per unit of crop yield based on Equation 11 (IPCC, 2014):

where GHGI is the GHG emission intensity (kg CO₂-eq t^-¹) and Y is the alfalfa hay yield (t ha^-¹).

After data processing with Excel 2016 (Microsoft Corp., Redmond, WA, USA), statistical analyses were performed using IBM SPSS Statistics 20.0 (IBM Corp., Armonk, NY, USA). Two-way analysis of variance (ANOVA) followed by the least significant difference (LSD) test was used to identify differences in cumulative GHG emissions, crop yield, IWP, PFPN, GHGI, and GWP among treatments. Pearson correlation and linear regression were applied to assess relationships between GHG fluxes and soil properties. Graphs were generated using Origin 2023 (OriginLab Corp., Northampton, MA, USA). Prior to ANOVA, normality of residuals (Shapiro–Wilk test) and homogeneity of variances (Levene’s test) were examined. All variables met both assumptions (P > 0.05).

During the experimental period, the daily maximum and minimum temperatures in the study area were 39.9°C and –10.1°C, respectively, with the total precipitation reaching 578.7 mm. The average daily temperature over the alfalfa growing season ranged from –6.7°C to 31.6°C (Figure 2). Soil temperatures (0–5 cm) exhibited consistent trends across treatments and showed clear seasonal variation, primarily driven by air temperature and precipitation (Figure 3a). Under different irrigation treatments, the average soil temperature for W0 (21.35°C) was slightly, but not significantly, higher than that of W1 (20.94°C) and lower than that of W2 (21.64°C). WFPS varied prominently due to irrigation and rainfall, showing repeated wet-dry cycles. WFPS increased with higher irrigation volumes, ranging between 24.52–75.23% for W0, 27.82–77.08% for W1, and 31.02–83.06% for W2. Within each irrigation cycle, WFPS gradually decreased over time in similar patterns (Figure 3b). Under different nitrogen treatments, soil available nitrogen contents (0–20 cm) displayed comparable seasonal changes, with temporary increases followed by declines after irrigation or rainfall. The sum of soil NO₃^– and NH₄⁺ contents ranged from 13.7 to 232.18 mg kg^-¹, and higher irrigation volumes or nitrogen rates maintained elevated soil nitrogen levels for longer durations. Soil available nitrogen contents trended downward over the alfalfa growing season (Figure 3c).

The relationships between GHG emission fluxes and soil physicochemical properties were assessed using correlation analysis. During the alfalfa growing season, N₂O fluxes linearly increased with rising soil temperature and available nitrogen content, and higher fluxes were observed within the WFPS range of 50–60%. CO₂ fluxes peaked at soil temperatures between 15°C–20°C and decreased linearly with increasing WFPS, with higher fluxes at WFPS of 35–45%. CH₄ fluxes were negatively correlated with soil temperature and positively correlated with soil WFPS and available nitrogen content. Further, the relative contributions of soil physicochemical properties to N₂O, CO₂, and CH₄ emission fluxes were assessed using linear regression. Regression equations were established for each gas emission flux against soil temperature, WFPS, and available nitrogen content. WFPS did not show significant linear correlations with N₂O, CO₂ or CH₄ fluxes (P > 0.05; Figure 4), probably because soil moisture rarely reached the critical thresholds for sharp emission pulses under the semi-arid conditions of this study, leaving mineral N as the dominant short-term driver.

N₂O fluxes frequently rose within a week of N application or ≥10 mm rainfall and then subsided, with the largest pulses in W2N3 (Figures 5a–c). Because background variability was high and the routine sampling interval was 7–10 d, we confined our inference to treatment-level cumulative emissions rather than to exact lag times. W2 resulted in the greatest fluctuation in N₂O fluxes over the alfalfa growing season (Figure 5c), followed by W1 (Figure 5b), and the smallest variation was observed for W0 (Figure 5a). Among the three harvest cycles, N₂O fluxes were highest in the second cycle (1.09–469.35 μg m^-² h^-¹) compared to the first (1.09–326.24 μg m^-² h^-¹) and third (5.01–199.68 μg m^-² h^-¹) cycles under each irrigation level. In the third harvest cycle, N₂O fluxes exhibited similar trends between W0 and W1, both of which were significantly lower than W2. Under different nitrogen treatments, N₂O fluxes trended upward with higher nitrogen rates. In the third harvest cycle without topdressing, there were significantly lower N₂O fluxes across all nitrogen treatments than in the first and second cycles. The average N₂O flux peaked in W2N3, and was lowest in W0N0.

Cumulative N₂O emissions were significantly affected by fertilization and its interactions with irrigation (Figures 6a–c). Combined irrigation and nitrogen application significantly increased cumulative N₂O emissions over the alfalfa growing season, and emissions increased proportionally with higher treatment levels. ANOVA results showed significant differences in cumulative N₂O emissions among nitrogen treatments, but not among irrigation treatments (Figures 6a, b). During the three harvest cycles, cumulative N₂O emissions decreased with lower nitrogen rates across all irrigation treatments and were lowest in the third cycle, with the highest in the first cycle. Cumulative N₂O emissions peaked in W2 (2.72 kg ha^-¹) and N3 (3.63 kg ha^-¹) (Figures 6a, b). Among the combined treatments, increasing nitrogen rate under each irrigation level elevated cumulative N₂O emissions. W2N3 produced the highest cumulative N₂O emissions (4.59 kg ha^-¹), with the lowest observed for W0N0 (0.89 kg ha^-¹) (Figure 6c). These results indicate that post-fertilization N₂O emissions were a major contributor to annual totals. In particular, emissions following basal and secondary nitrogen applications accounted for ~53% of the annual cumulative N₂O emission.

With regard to CO₂ emission fluxes, there were significant effects of irrigation, fertilization, and their interactions (Figures 5d–f). CO₂ fluxes showed consistent patterns across treatments, with W2 exceeding W1 and N1–N3 slightly higher than N0. Distinct temporal patterns emerged in CO₂ fluxes under different water-nitrogen interactions. After fertilization and the first irrigation (March 14), maximum CO₂ flux occurred within the first week of input, but equally rapid rises followed subsequent rains; thus, timing is co-controlled by temperature and moisture, not a fixed lag. Subsequent irrigation and rainfall events triggered new peaks in CO₂ fluxes. After fertilization and the second irrigation (April 14), W0N0 maintained relatively stable CO₂ fluxes, which peaked at 1033.76 mg m^-² h^-¹ on day 6 and then decreased afterwards. Following rainfall, CO₂ fluxes remained persistently high for 0–6 days. Additionally, CO₂ fluxes were temperature-controlled, generally lower in the cold months (November–March) than in the warm months (April–October).

Changes in cumulative CO₂ emissions indicated significant effects of irrigation and irrigation-fertilization interactions (Figures 6d–f). ANOVA results revealed that cumulative CO₂ emissions differed significantly among irrigation treatments, but not among nitrogen treatments (Figures 6d, e). Specifically, cumulative CO₂ emissions increased with higher irrigation volumes, reaching 22,586.17 kg ha^-¹ for W2. Within each irrigation treatment, cumulative CO₂ emissions exhibited a mild upward trend with increasing nitrogen rate. The lowest annual cumulative CO₂ emission was observed for W0N0 (10,530.18 kg ha^-¹) and the highest for W2N3 (25,206.11 kg ha^-¹) (Figure 6f). Emissions following irrigation and fertilization accounted for ~1.56% of the annual cumulative CO₂ emissions across treatments.

CH₄ emission fluxes were significantly responsive to irrigation, but not to fertilization or their interactions (Figures 5g–i). All field plots consistently functioned as CH₄ sinks throughout the experimental period. Soil CH₄ uptake was transiently reduced after wetting, yet the response was inconsistent; only irrigation level significantly affected the seasonal sink strength. CH₄ fluxes decreased with increasing irrigation volume, ranging from –77.95 μg m^-² h^-¹ (W0) to –8.14 μg m^-² h^-¹ (W2). No distinct temporal pattern in CH₄ fluxes was observed across treatments. The effects of irrigation on cumulative CH₄ uptake were also significant (Figures 6g–i). Cumulative CH₄ uptake decreased with increasing irrigation volume, from 1.50 kg ha^-¹ (W0) to 0.60 kg ha^-¹ (W2) (Figure 6h). Emission peaks following irrigation and fertilization contributed ~1.02% to the annual cumulative CH₄ emission.

Irrigation, fertilization, and their interactions all exhibited highly significant effects on alfalfa hay yield (Figure 7). Higher irrigation volumes led to improved alfalfa yields. Across all irrigation treatments, alfalfa yields increased initially and then decreased over successive harvest cycles. The mean yield was highest in the second harvest cycle (7,811.0 kg ha^-¹) followed by the third cycle (5,232.4 kg ha^-¹), and the lowest yield was recorded for the first cycle (3,799.1 kg ha^-¹). Under W0 and W1 treatments, alfalfa yield increased with higher nitrogen rates, peaking at 6,069.46 kg ha^-¹ for N3. Under W2 treatment, alfalfa yield peaked at 6,876.73 kg ha^-¹ for N2 and then declined with higher nitrogen input; N3 resulted in lower alfalfa yield than N2 and had no significant difference compared to N1. The annual yield across irrigation-nitrogen treatments ranged from 18,907.30 kg ha^-¹ (W2N2) to 11,811.79 kg ha^-¹ (W0N0).

The effects of irrigation, fertilization, and their interactions on IWP were highly significant (Table 2). IWP was lower under W2 than under W1; however, this comparison is restricted to only two irrigated levels (W1 and W2) and should therefore be treated as indicative rather than a statistically robust trend. Across nitrogen treatments, IWP initially increased and then declined with increasing N rate. Among different nitrogen treatments, the highest IWP was observed for N2, exceeding that of N0 by 26.2% (W1) and 39.6% (W2). Under the combined treatments, IWP peaked in W1N3 (6.82 kg m^-³), followed by W1N2 (6.67 kg m^-³) in the second harvest cycle. W2N2 improved IWP by 10.9% compared to the conventional high-water, high-nitrogen regime W2N3.

The PFPN patterns also indicated highly significant effects of irrigation, fertilization, and their interactions (Table 2). PFPN increased with higher irrigation volumes, while it negatively responded to nitrogen rate under each irrigation level. Among all combined treatments, the highest PFPN was recorded for W2N1 (300.55 kg kg^-¹), and the lowest for W0N3 (94.87 kg kg^-¹). W2N2 achieved a favorable balance between alfalfa yield and nitrogen use efficiency, improving PFPN by 56.9% compared to W2N3.

To assess the climate response to GHG emissions in alfalfa fields under water-nitrogen management, ANOVA was conducted for EF, GWP, and GHGI across treatments. All three parameters showed significant differences in relation to irrigation, fertilization, and their interactions (Table 3). EF showed an upward trend with increasing irrigation volume. Under both W0 and W2 levels, EF rose with higher nitrogen rates. Under the W1 level, N2 produced the lowest EF (1.10%). The highest EF value (1.86%) was observed for W2N3, which also recorded the greatest GWP and GHGI values under combined treatments. When incorporating indirect emissions from agricultural inputs, the total GWP for W2N3 reached 3721.54 kg CO₂-eq ha^-¹. Compared to W2N3, W2N2 achieved favorable trade-offs, reducing EF (by 29.1%),total GWP (by 24.1%), and GHGI (by 29.0%). This reduction was primarily driven by lower nitrogen fertilizer inputs, mitigating both direct soil emissions and upstream industrial carbon costs. In summary, soil GHG emissions showed a strong positive response to high water and nitrogen inputs, while the high-water, medium-nitrogen regime effectively reduced GHG emissions from alfalfa production. The highest N₂O emission factor (1.86%) under W2N3 approached twice the IPCC default (1.0%), while optimized treatments (≤ 120 kg N ha^-¹) fell within ranges reported for alfalfa systems (0.4–1.7%; Fernández-Ortega et al., 2024; Kamran et al., 2022).

N₂O is produced mainly through microbial nitrification and denitrification processes, and soil N₂O emissions are highly sensitive to soil moisture content, temperature, and nitrogen availability (Qian et al., 2025). Irrigation can indirectly promote N₂O emission fluxes. In the absence of nitrogen application, increased irrigation levels had a limited effect on soil N₂O emissions from alfalfa fields. However, when nitrogen was applied, irrigation markedly enhanced N₂O emissions, possibly because increased soil moisture content indirectly contributed to N₂O emissions by improving NO₃^–N availability (Ullah et al., 2016). It has been reported that elevated WFPS levels are associated with enhanced soil denitrification (Anderson et al., 2019). N₂O emission fluxes were promoted by soil WFPS, as indicated by their significant positive correlation observed in this study. The underlying mechanism is that high irrigation levels (WFPS 50–60%) facilitated soil denitrification. This finding is consistent with the denitrification thresholds reported by Bracken et al. (2021) and Wang et al. (2023).

The application of nitrogen fertilizer resulted in considerably higher soil N₂O emissions in alfalfa fields, as nitrogen application elevated soil available nitrogen content, providing ample substrates for nitrifying and denitrifying bacteria (Xun et al., 2022). N₂O emission fluxes were positively driven by soil available nitrogen content, based on the results of correlation and regression analyses. There were stepwise rises in N₂O fluxes from low nitrogen (N1) to high nitrogen (N3) treatments compared to the unfertilized control (N0), and the fluxes had strong positive correlation with soil available nitrogen content that surged immediately after fertilization. NO₃^–N is considered the primary source of nitrogen for N₂O emissions. Excessive nitrogen not absorbed by alfalfa roots remains in the soil, serving as substrates for microbial nitrification and denitrification, consequently stimulating N₂O emissions (Jiang et al., 2025; Wang et al., 2024). These findings emphasize the importance of optimizing irrigation and nitrogen rates to control soil N₂O emissions from alfalfa cultivation. Given the 7–10 d routine sampling interval, we confine our interpretation to treatment-level cumulative emissions (Figures 6a–c) rather than to exact lag times.

Soil CO₂ fluxes in alfalfa fields reflected combined microbial and root respiration, with a smaller contribution from carbonate dissolution (Bai et al., 2021). Nitrogen fertilization stimulated microbial mineralization and root metabolic activity, thereby increasing CO₂ release during N transformation processes (Hauck, 1981; Snyder et al., 2009; Kong et al., 2020). Short-term CO₂ pulses appeared immediately after irrigation and fertilization, consistent with observations under fixed-moisture conditions (Pan et al., 2023). However, the strength of these responses differed among growth stages, indicating that water-nitrogen (W-N) interactions do not exert a uniform influence throughout the season.

The interaction was strongest during early regrowth. Following irrigation, CO₂ fluxes increased sharply, reflecting the rewetting-induced activation of soil microbes documented in dryland systems (Zhai et al., 2011; Badaou and Sahin, 2025). These pulses overlapped with nitrogen addition, which provided readily available substrates for microbial decomposition and root respiration, producing higher peaks than those reported in single-factor experiments (Pan et al., 2023). The highest fluxes occurred on days 7 and 6 after irrigation, highlighting the system’s sensitivity to combined moisture and nutrient inputs. During peak biomass, CO₂ fluxes were driven mainly by temperature and sustained soil moisture, whereas nitrogen effects weakened. Irrigation maintained higher soil moisture and supported root and microbial respiration, resulting in notably greater cumulative CO₂ emissions under W2 than W0. This pattern aligns with evidence that adequate moisture enhances autotrophic and heterotrophic respiration when plant growth is vigorous (Wang et al., 2022). After harvest, CO₂ fluxes declined and W-N interactions were minimal. Reduced root activity and lower belowground carbon allocation (Kong et al., 2020) diminished nitrogen-induced differences, while irrigation produced only moderate variation through soil moisture regulation.

Overall, W-N interactions had their strongest effects immediately after irrigation during early regrowth, whereas temperature and sustained moisture largely controlled cumulative CO₂ emissions. This stage-dependent behavior explains the varying responses among harvest periods and underscores the combined influence of moisture pulses and plant phenology on CO₂ dynamics in semi-arid alfalfa systems.

CH₄ is produced by methanogenic archaea through anaerobic decomposition of soil organic matter. Under aerobic conditions, methanotrophic microorganisms consume CH₄ as an energy source along with O₂. Therefore, the soil functions as a CH₄ source when CH₄ production by methanogens exceeds its oxidation by methanotrophs, and shifts to a CH₄ sink when methanotrophs become dominant (Angle et al., 2017). During alfalfa cultivation, different water-nitrogen regimes did not significantly affect CH₄ fluxes, supporting the results from earlier pot and field experiments showing that alfalfa soils consistently act as CH₄ sinks (Fernández-Ortega et al., 2024; Kamran et al., 2022; Ning et al., 2020). However, previous studies collected monthly data from static chamber measurements or considered only one irrigation level, which contrasts with our high-frequency (daily) measurements across three irrigation and four nitrogen levels. The results showed that even pronounced wet-dry cycles and nitrogen pulses did not stimulate CH₄ uptake or cause net emissions, highlighting the robustness of the soil sink function under intensive management. Based on the positive correlation between soil CH₄ emission fluxes and WFPS, high WFPS levels under intensive irrigation likely improved soil porosity and structure, creating anaerobic conditions favorable for methanogenic activity. Similarly, Chen et al. (2021) reported that high soil moisture content reduced O₂ availability, expanding anaerobic zones and enhancing CH₄ production by methanogens. Although WFPS showed no significant linear correlation with CH₄ fluxes (Figure 4), non-linear or threshold moisture effects may exist; these could not be tested with the present single-year dataset, so only significant linear relationships are retained in Figure 4 and the complete matrix is provided in Supplementary Image 1.

The effect of nitrogen application on CH₄ emissions remains inconclusive. Nitrogen rate did not significantly affect CH₄ emissions from alfalfa production, although findings from other crops may offer useful insights. For example, nitrogen fertilization has been identified as a major factor affecting CH₄ emissions in cotton systems (Bai et al., 2021). In wheat systems, moderate nitrogen levels may enhance CH₄ oxidation, while excessive fertilization or topdressing can inhibit methanotrophic activity, resulting in increased CH₄ emissions (Hütsch et al., 1993).

Irrigation and fertilization play essential roles in achieving high yield and quality of alfalfa in arid and semi-arid regions (Ma et al., 2024). While nitrogen application increased annual alfalfa yield compared to the control, there was a non-linear yield response to nitrogen rate. This pattern is attributed to the role of nitrogen as a key determinant of crop productivity. When applied at optimal levels, nitrogen improves alfalfa yield by promoting cell division and growth and enhancing dry matter accumulation (Lv et al., 2025; Ma et al., 2024). However, excessive nitrogen input may suppress rhizobial nitrogen fixation, reducing nitrogen use efficiency (Hu et al., 2019). Our results indicated that high nitrogen application (N3) did not improve alfalfa yield, potentially due to reduced nitrogen use efficiency and increased GHG emissions. Additionally, alfalfa yield increased progressively with higher irrigation levels, consistent with previous findings by Jia et al. (2018) under a single nitrogen level. The rate of yield increase was amplified under moderate nitrogen application and peaked when the nitrogen level reached 120 kg ha^-¹. During the alfalfa growing season, substantial rainfall in June–August 2024 provided sufficient soil moisture, leading to the highest crop yield in the second harvest. It has been suggested that moderately increasing fertilization and irrigation promotes improvements in fresh forage yield, while over-fertilization with poor irrigation management can lead to yield reduction (Kandelous et al., 2012). No yield reduction due to irrigation was observed in the present study, possibly because the applied water volumes did not exceed the critical threshold. Within a suitable soil moisture range, appropriate fertilization notably improved alfalfa productivity (Mu et al., 2025).

IWP and PFPN are key indicators for evaluating water and fertilizer management in agricultural systems, and both of them are strongly affected by irrigation and fertilization strategies. Wen et al. (2024) suggested that IWP increases with higher irrigation volumes within a specific range, whereas Djaman et al. (2020) reported the opposite pattern. In this study, IWP decreased as irrigation volume increased, and increasing nitrogen application improved IWP under each irrigation level, suggesting a synergistic interaction between water and fertilizer inputs. Appropriate fertilization can promote root growth and distribution, enhancing water and nutrient uptake, thereby improving IWP under water-limited conditions. Additionally, it has been shown that PFPN increases first and then declines with increasing nitrogen rate over a reasonable range (Mu et al., 2025). This suggests that moderate nitrogen input enhances crop yield, whereas excessive nitrogen application causes fertilizer loss and environmental risk. Results of this study showed that PFPN declined with increasing nitrogen rate across all irrigation levels, indicating that excessive nitrogen input did not necessarily improve fertilizer use efficiency. Coordinated water and fertilizer management during alfalfa cultivation achieved the dual goals of “managing water through fertilizer” and “enhancing fertilizer use efficiency through water.” The water-fertilizer synergy was maximized under combined application of appropriate irrigation and nitrogen levels, effectively improving alfalfa yield and resource use efficiency. For example, both PFPN and IWP peaked under moderate nitrogen application (180 kg ha^-¹) with high irrigation (300 mm).

Soil GHG emissions remained positive across all irrigation and nitrogen treatments, indicating that the alfalfa system acted as a net GHG source. GWP increased under higher irrigation levels with elevated nitrogen rates, suggesting that both water and nitrogen were key drivers influencing this metric. Specifically, the high GWP in intensive treatments was driven not only by soil denitrification but also by the substantial carbon footprint embedded in agricultural inputs. N₂O and CH₄ are recognized as the two most important soil-derived GHGs in GWP assessments (Li et al., 2023), and their emissions are highly sensitive to both irrigation and fertilization practices. Under arid soil conditions, CH₄ uptake often exceeds its emissions, making N₂O the dominant contributor to soil-based GWP in alfalfa systems (Zhang et al., 2024). Fertilizer application can promote microbial activity in the soil. For example, NH₄⁺-based nitrogen fertilizer provides readily available substrate for ammonia-oxidizing bacteria, stimulating soil nitrification (Fudjoe et al., 2023). However, excessive application of basal or topdressing fertilizer disrupts the balance between soil nitrogen supply and plant uptake. This could result in residual mineral nitrogen and elevated N₂O emissions, consequently increasing soil GWP.

Although this study integrated LCA to calculate comprehensive GWP, it should be noted that excluding soil CO₂ introduces uncertainty because chamber-based CO₂ respiration does not represent net ecosystem carbon exchange. Therefore, the GWP and GHGI values reported here reflect only direct non-CO₂ soil fluxes and indirect input emissions, explicitly excluding chamber-measured soil respiration. Future studies should combine chamber measurements with ecosystem-scale carbon assessments for a more complete evaluation. This approach aligns with protocols focusing on non-CO₂ greenhouse gases in short-term studies (Shang et al., 2019; Smith et al., 2020).

GHGI is a key metric for evaluating agricultural sustainability. Optimal fertilization can enhance alfalfa yield while lowering GHGI (Ma et al., 2024). This study showed that moderate nitrogen application (N2) resulted in higher alfalfa yield (by 1.15–7.02%) and lower LCA-based GHGI (by ~29.0%) compared to high nitrogen treatment (N3). The reduction in GHGI for N2 was due to the dual mitigation of direct soil N₂O emissions and the indirect carbon footprint from fertilizer production. Our LCA analysis revealed that ‘hidden’ carbon emissions from inputs are major drivers of GWP; optimizing the N rate to 120 kg ha^-1 (W2N2) effectively cut these upstream emissions by approximately 33% compared to the N3 level. Therefore, alleviating GHGI in agricultural systems requires optimized nitrogen management to lower GHG emissions and improve crop yields—dual objectives that can be simultaneously achieved by adjusting the fertilization rate. Additionally, rational irrigation management can reduce nutrient leaching and GHG emissions associated with excessive water use (Wang et al., 2025).

Aligning irrigation with crop water demand is helpful to minimize GHGI while sustaining crop productivity. Conventional practices involving high water and nitrogen inputs were associated with elevated GHGI, whereas the optimal treatment (W2N2) reduced environmental impacts, resulting in lower GHGI. This study covered only one growing season at a single site; both weather conditions and GHG emissions are known to exhibit substantial inter-annual variability. Consequently, the proposed water–nitrogen management recommendations should be viewed as site-specific and year-specific, rather than universally optimized thresholds. Multi-year, multi-location experiments are needed to test the stability and generality of the observed synergy zone.