Effects of Long-Term Controlled-Release Urea on Soil Greenhouse Gas Emissions in an Open-Field Lettuce System

Controlled-release urea (CRU) fertilizers are widely used in agricultural production to reduce conventional nitrogen (N) fertilization-induced agricultural greenhouse gas emissions (GHGs) and improve N use efficiency (NUE). However, the long-term effects of different CRU fertilizers on GHGs and crop yields in vegetable fields remain relatively unexplored. This study investigated the variations in GHG emissions at four growth stages of lettuce in the spring and autumn seasons based on a five-year field experiment in the North China Plain. Four treatments were setup: CK (without N application), U (conventional urea—N application), ON (20% reduction in urea—N application), CRU (20% reduction in polyurethane-coated urea without topdressing), and DCRU (20% reduction in polyurethane-coated urea containing dicyandiamide [DCD] without topdressing). The results show that N application treatments significantly increased the GHG emissions and the lettuce yield and net yield, and DCRU exhibited the lowest N2O and CO2 emissions, the highest lettuce yield and net yield, and the highest lettuce N content of the N application treatments. When compared to U, the N2O emission peak under CRU and DCRU treatments was notably decreased and delayed, and their average N2O emission fluxes were significantly reduced by 10.20–20.72% and 17.51–29.35%, respectively, leading to a significant reduction in mean cumulative N2O emissions during the 2017–2021 period. When compared to U, the CO2 fluxes of DCRU significantly decreased by 8.0–16.54% in the seedling period, and mean cumulative CO2 emission decreased by 9.28%. Moreover, compared to U, the global warming potential (GWP) and greenhouse gas intensity (GHGI) of the DCRU treatment was significantly alleviated by 9.02–17.13% and 16.68–20.36%, respectively. Compared to U, the N content of lettuce under DCRU was significantly increased by 6.48–17.25%, and the lettuce net yield was also significantly increased by 5.41–7.71%. These observations indicated that the simple and efficient N management strategy to strike a balance between enhancing lettuce yields and reduce GHG emissions in open-field lettuce fields could be obtained by applying controlled-release urea containing DCD without topdressing.


Introduction
As an important part of the terrestrial ecosystem, agroecosystems are a major source of global greenhouse gas (GHG) emissions [1], including nitrous oxide (N 2 O), methane (CH 4 ), and carbon dioxide (CO 2 ), accounting for 14-17% of global directly anthropogenic GHG emissions [2][3][4].The contribution of these GHGs depends on their global warming potentials and atmospheric lifespans.Specifically, CO 2 is the leading contributor to climate change, while non-CO 2 GHGs (N 2 O and CH 4 ) also play a crucial role despite their lower emissions.N 2 O has 298 times the global warming potential of CO 2 based on a 100-year timescale [2,5].GHG emissions from croplands depend on both environmental variations associated with climate and soil type and agricultural management such as tillage, fertilization and irrigation [6,7].The excessive use of nitrogen (N) fertilizer has improved crop yields [8], but has also become the largest contributor of GHGs from farmland, accounting for more than 25% of total global N 2 O emissions [9,10].Hence, there exists an urgent need to develop optimized N fertilization strategies that can mitigate GHG emissions without detrimental impacts on crop yields [5,11].
Many studies have suggested that controlled-release urea (CRU), a newly developed long-acting N fertilizer, could be adopted as an optimal N management practice to improve N use efficiency (NUE), increase crop yields, and reduce GHG emissions.CRU can better match fertilizer nutrient release with crop uptake without additional topdressing [10,12,13].Previous meta-analyses indicated that CRU could be applied once as basal fertilizer with no effect on grain yields while reducing reactive N losses by 49% worldwide [14], thus decreasing the use of N fertilizer and promoting time-and labor-saving crop production [15,16].The advantageous effects of CRU on grain yield and NUE depend largely on the synchronization of the N release rate with the N requirements of crops [10].The application of CRU slows down the release of available N through its physical coating, influences soil NH 4 + -N and NO 3 − -N content and C/N ratio [17], and alters the structure and activity of the soil microbial community [18].This may affect soil C and N cycling and then alter CO 2 and/or reduce N 2 O emissions [12,18].Although the coated material can control the release rate of nutrients, it cannot control the transformation of dissolved urea in soil, thereby still causing some N losses [13].
The combined application of an nitrification inhibitor (NI) and urea is another way to reduce N 2 O emissions, which can control N transformation in soil and improve the NUE [19][20][21].Dicyandiamide (DCD) is the most widely used NI with relatively low volatility.Previous studies demonstrated that DCD reduced the N 2 O emission by suppressing nitrification and denitrification in soil [19,[22][23][24].Additionally, a three-year field experiment indicated that the application of DCD reduced CO 2 emissions by 7% [25].Ahmed et al. [3] reported that urea applied with 1.5% DCD decreased the CO 2 and N 2 O emissions in soil by 22-172% and up to 94-fold, respectively, compared with those under conventional N fertilizer application.However, the performance and stability of DCD and the transformation of dissolved urea released from CRU vary with soil properties such as soil temperature, pH, and water content [26,27].Fan et al. [19] confirmed that the CRU amended with 3,4-dimethylpyrazole phosphate (DMPP) could reduce the GHG emissions of open-field vegetables, and that the mitigation effect was made more significant by combining CRU with DMPP.Ge et al. [27] confirmed that the application of CRU and DCD led to significant improvement in wheat yield and NUE, the effect of which was more pronounced under the synergistic use of resin-coated urea and NI.
Vegetable production usually involves intensive fertilization with high N loss potential and contributes to 9% of global cropland N 2 O emissions [28].China is the largest producer of vegetables in the world, accounting for 41% of total global production [29].Intensive vegetable production systems take up 13% of the Chinese crop planting area and consume 25% of the national production of chemical fertilizers while emitting 35% of crop-sourced GHGs due to the excessive N (inorganic and organic) inputs, multicropping, and frequent irrigation [30,31].Therefore, it is imperative to develop appropriate strategies to mitigate GHG emissions from vegetable fields.Some studies have indicated that optimizing N fertilizer management could significantly reduce GHG emissions and maintain the high vegetable yields of Chinese vegetable production systems [13,31].However, the combined effects of CRU and DCD and the effects of coated urea containing DCD on the GHG emissions from vegetable production systems remain unknown.Lettuce is widely planted in northern China, and mainly cultivated in open fields during spring and autumn.The impacts of optimized N fertilizer management on GHG emissions from open-field lettuce production demand in-depth exploration.
Therefore, a five-year field experiment was conducted to investigate the GHG emissions from an open-field lettuce system fertilized with two types of CRU in northern China.The objectives of this study were to: (1) investigate the long-term effects of different N fertilization treatments on GHG emissions (i.e., N 2 O and CO 2 ), soil available N contents, lettuce yield, and N uptake; (2) elucidate the underlying mechanisms and provide guidance for local farmers to optimize N fertilization strategies for the sustainable production of open-field vegetables.

N 2 O and CO 2 Emissions
Compared with CK, soil N 2 O emission fluxes increased significantly in N application treatments.The N fertilization treatment significantly affected soil N 2 O emission fluxes from lettuce soil, without significant inter-annual or inter-seasonal differences (Table 1 1).The average N 2 O fluxes reached their lowest level during the mature period, and those in spring were higher than autumn (Figure 1).However, the N 2 O fluxes of U and ON were at their highest at the seedling stage, with those in autumn higher than in spring.The average N 2 O fluxes of CRU and DCRU at the lotus stage were higher than those of other periods in spring, similar to the seedling stage in autumn.The results indicate that the long-term use of controlled-release fertilizers had a significant effect on N 2 O reduction without significant inter-annual or inter-seasonal differences.
Compared with CK, soil CO 2 emission fluxes were significantly increased in the N application treatments.The CRU and DCRU reduced the CO 2 emission fluxes from lettuce soil (p > 0.05); however, there were no significant differences between different years and seasons (Table 2).In 2017 to 2021, the average CO 2 fluxes were 347.99-362.82mg m −2 h −1 under different N fertilization treatments during the autumn, i.e., lower than those of the spring (Table 2).The average CO 2 fluxes gradually decreased from the seedling to the harvest stage in autumn, and those of CRU and DCRU at the lotus stage were higher than those of other periods in spring.The CO 2 fluxes from seedlings in autumn was higher than that in spring (Figure 2).During the seedling period, compared with U, the CO 2 fluxes of DCRU were reduced by 8.0-16.54%(Figure 2).The results indicate that the long-term use of DCRU fertilizer could effectively reduce CO 2 emissions.
The cumulative emissions of N 2 O and CO 2 were consistent with the patterns of N 2 O and CO 2 fluxes (Figure 3a,b).The cumulative emissions of N 2 O and CO 2 were 3.95-7.51kg ha −1 and 9127.18-11003.84kg ha −1 in the N application treatments, respectively, and decreased under CRU and DCRU treatments in the 2017-2021 period, achieving the lowest values under DCRU.Notably, compared with U, the cumulative emissions of N 2 O from ON, CRU, and DCRU were significantly reduced by 15.70%, 25.34%, and 37.68% (p < 0.05), and CO 2 by 4.46%, 6.77%, and 9.28% (p > 0.05), respectively (Figure 3a,b).The GWP was 10.66-13.05Mg CO 2 -eq ha −1 yr −1 in this research, compared to U, which decreased by 1.62-7.82%,6.76-11.85%,and 9.02-17.13%under ON, CRU, and DCRU, respectively, and the GWP of DCRU was significantly reduced by 13.91% (p < 0.05) from 2017 to 2021 (Figure 3c).The GHGI was 0.078-0.109 in this research, compared to U, which decreased by 5.06-11.10%,9.47-17.79%,and 16.68-20.17%in ON, CRU, and DCRU, respectively.Compared with U, the GHGI of DCRU from 2017 to 2021 was significantly reduced by 20.36% (p < 0.05) (Figure 3d).The results indicate that the long-term use of DCRU had a significant mitigation effect on the cumulative N 2 O emissions.local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.Compared with CK, soil CO2 emission fluxes were significantly increased in the N application treatments.The CRU and DCRU reduced the CO2 emission fluxes from lettuce soil (p > 0.05); however, there were no significant differences between different years and seasons (Table 2).In 2017 to 2021, the average CO2 fluxes were 347.99-362.82mg m −2 h −1 under different N fertilization treatments during the autumn, i.e., lower than those of the spring (Table 2).The average CO2 fluxes gradually decreased from the seedling to the harvest stage in autumn, and those of CRU and DCRU at the lotus stage were higher than those of other periods in spring.The CO2 fluxes from seedlings in autumn was higher than that in spring (Figure 2).During the seedling period, compared with U, the CO2 fluxes of DCRU were reduced by 8.0-16.54%(Figure 2).The results indicate that the long-term use of DCRU fertilizer could effectively reduce CO2 emissions.

Lettuce Yield and N Content
Compared with CK, the lettuce yield and net yield was significantly increased in N application treatments (Figure 5a-d).Compared with CK, the lettuce yield of ON, CRU and DCRU treatments slightly increased with no significant differences (p > 0.05) (Figure 5a,b).Compared with U, the net yield of DCRU was significantly increased by 7.47%, 5.83%, and 5.41% (p < 0.05) during the spring season in 2019, 2020, and 2021, respectively, and by 6.79%, 7.48%, 7.35%, and 7.71% during the autumn season in 2017, 2018, 2019, and 2021, respectively (Figure 5c,d).The results indicated that the long-term use of DCRU fertilizer increased the net yield.Compared with CK, the N content of lettuce was significantly increased in application treatments (Figure 6a,b).The N content of lettuce showed no significa differences across different N treatments, seasons, and years (Figure 6a,b).Compar with U, the N content of spring lettuce in ON, CRU, and DCRU was increased by 3.5 4.67%, 5.27-7.63%,and 7.35-12.20%,and that of autumn lettuce was increased by 3.6 8.82%, 3.72-9.96%,and 6.48-17.25%.The results indicated that the long-term use DCRU could effectively increase the N content of lettuce.Compared with CK, the N content of lettuce was significantly increased in N application treatments (Figure 6a,b).The N content of lettuce showed no significant differences across different N treatments, seasons, and years (Figure 6a,b).Compared with U, the N content of spring lettuce in ON, CRU, and DCRU was increased by 3.51-4.67%,5.27-7.63%,and 7.35-12.20%,and that of autumn lettuce was increased by 3.60-8.82%,3.72-9.96%,and 6.48-17.25%.The results indicated that the long-term use of DCRU could effectively increase the N content of lettuce.

Effects of Different Fertilization Treatments on N2O Emissions
The N2O in soil originate from the microbial nitrification processes under aerobic conditions and the denitrification processes under anaerobic conditions [32].The N2O emissions are affected by soil inorganic N content, soil temperature and water filled pore space (WFPS) [33,34].In general, vegetable soils with high chemical N fertilizer inputs and frequent irrigation are major sources of N2O [11,34,35].Wang et al. [31] conducted a meta-analysis focused on Chinese vegetable production systems and found that N2O emissions had a positively linear correlation with fertilizer N application rates.Our results were consistent with previous studies.Due to the large amount of N fertilization, the annual N2O emissions from the open-field lettuce system reached 7.51 kg ha −1 under conventional N application practices (spring and autumn, N 480) (Figure 3a) in this study, which were 1.41 times higher than a wheat-corn rotation system (3.12 kg ha −1 ) in the same region [36].In this study, cumulative N2O emissions were reduced by 15.70% under ON compared to U. Therefore, it is necessary to establish optimized N management strategies in open-field lettuce systems to decrease the N2O emissions in this region [7].
Our findings suggest that, compared to CK, the N fertilization treatments significantly affected N2O emissions.The CRU significantly decreased the soil N2O emissions from lettuce fields compared to conventional N fertilization (Figures 1 and 3), which was in accordance with previous findings [10,14,37].The soil NH4 + -N and NO3 − -N levels were significantly lower under CRU than ON, possibly because conventional urea dissolved rapidly in the soil solution upon entering the soil.This rapid dissolution leads to a short-term increase in soil N content, reducing N uptake due to limited lettuce growth.As a result, greater accumulation of extra N substrates promotes the nitrification and denitrification processes, ultimately leading to increased N2O production [38][39][40].In our study, the polyurethane film as the outer layer of CRU prevented water penetration into the fertilizer, thereby slowing down the release of N, promoting lettuce uptake of mineral N from the soil, increasing N fertilizer utilization and lettuce yield, reducing substrate N concentrations required for N2O formation, and suppressing soil nitrification and denitrification, ultimately resulting in decreased N2O emissions [36,41,42].Except for the high N2O fluxes under traditional urea fertilization in 2017-2021 due to the high soil NO3 − -N contents, the N2O fluxes of the seedling stage (basal

Effects of Different Fertilization Treatments on N 2 O Emissions
The N 2 O in soil originate from the microbial nitrification processes under aerobic conditions and the denitrification processes under anaerobic conditions [32].The N 2 O emissions are affected by soil inorganic N content, soil temperature and water filled pore space (WFPS) [33,34].In general, vegetable soils with high chemical N fertilizer inputs and frequent irrigation are major sources of N 2 O [11,34,35].Wang et al. [31] conducted a meta-analysis focused on Chinese vegetable production systems and found that N 2 O emissions had a positively linear correlation with fertilizer N application rates.Our results were consistent with previous studies.Due to the large amount of N fertilization, the annual N 2 O emissions from the open-field lettuce system reached 7.51 kg ha −1 under conventional N application practices (spring and autumn, N 480) (Figure 3a) in this study, which were 1.41 times higher than a wheat-corn rotation system (3.12 kg ha −1 ) in the same region [36].In this study, cumulative N 2 O emissions were reduced by 15.70% under ON compared to U. Therefore, it is necessary to establish optimized N management strategies in open-field lettuce systems to decrease the N 2 O emissions in this region [7].
Our findings suggest that, compared to CK, the N fertilization treatments significantly affected N 2 O emissions.The CRU significantly decreased the soil N 2 O emissions from lettuce fields compared to conventional N fertilization (Figures 1 and 3), which was in accordance with previous findings [10,14,37].The soil NH 4 + -N and NO 3 − -N levels were significantly lower under CRU than ON, possibly because conventional urea dissolved rapidly in the soil solution upon entering the soil.This rapid dissolution leads to a shortterm increase in soil N content, reducing N uptake due to limited lettuce growth.As a result, greater accumulation of extra N substrates promotes the nitrification and denitrification processes, ultimately leading to increased N 2 O production [38][39][40].In our study, the polyurethane film as the outer layer of CRU prevented water penetration into the fertilizer, thereby slowing down the release of N, promoting lettuce uptake of mineral N from the soil, increasing N fertilizer utilization and lettuce yield, reducing substrate N concentrations required for N 2 O formation, and suppressing soil nitrification and denitrification, ultimately resulting in decreased N 2 O emissions [36,41,42].Except for the high N 2 O fluxes under traditional urea fertilization in 2017-2021 due to the high soil NO 3 − -N contents, the N 2 O fluxes of the seedling stage (basal fertilization) were higher than those of the mature period.This was possibly because the basal fertilizer was applied with 40% of the total N inputs, and the lower N requirement for lettuce growth during the seedling stage, resulting in Plants 2024, 13, 1071 9 of 16 higher soil NH 4 + -N and NO 3 − -N contents during the seedling stage than the mature period, thus stimulating N 2 O emissions and leading to higher N 2 O fluxes than those from the mature period.Moreover, high concentration of NO 3 − may also inhibit the reduction of soil N 2 O to N 2 , which thus increases the soil N 2 O emissions under higher WFPS [43].
Our study found that the DCRU fertilizer outperformed other treatments in mitigating N 2 O emissions from the open-field lettuce system, possibly due to the synergistic effects of NIs and CRU fertilizers.Previous findings showed that NI amendments are an effective approach to decrease direct soil N 2 O emissions [19,21,44].The inhibition of the abundance of nitrifiers by NIs directly reduced the nitrification rate, and lowered the soil NO 3 − availability as the substrate for denitrification [13], leading to a reduction in direct N 2 O emissions.Our study also found that compared with conventional N fertilization, the DCRU treatment decreased soil NO 3 − -N by 13.78-21.59%,which probably led to the decrease in N 2 O (17.51-29.35%).
In addition to fertilization, other factors such as the irrigation amount [30] and soil properties [11] have also been reported to significantly affect N 2 O emissions in vegetable production systems.Our study found that frequent irrigation and precipitation during the two lettuce seasons (May and June in spring and August and September in autumn) resulted in high soil WFPS (50-73%) and temperature (mean: 13.28-24.58• C) (Table S1).This was beneficial to N 2 O generation, which was mostly produced via denitrification and oxidizer-denitrification [30,34].The N 2 O fluxes during the seedling stage in autumn were higher than those in the spring season (Figure 1), due to higher soil temperature and moisture during the seedling stage in autumn and the stimulated soil microbial activity involved in the N cycle, subsequently promoting soil N 2 O production [35].CRU controls the release rate of N based on soil moisture and temperature [45].The N 2 O fluxes of CRU and DCRU at the lotus stage in spring and the seedling stage in autumn were higher than those of the other periods, which was possibly because the soil moisture and temperature at this stage favored the rapid release of N and enhanced the N substrate for N 2 O production.

Effects of Different Fertilization Treatments on CO 2 Emissions
The results of this study indicated that N fertilization practices imposed a significant impact on CO 2 emissions, with the highest being observed from conventional N fertilization.Our results are consistent with previous studies [46,47].Huang et al. [48] found that CO 2 emissions under urea fertilized treatments were two-fold higher compared to non-fertilized treatments.Zhong et al. [47] showed that soil CO 2 emissions in farmland ecosystems increased with the N application rate.Lin et al. [49] found that, over six growing seasons, the mean growing season soil C emissions were increased by 11.6-29.7%under N fertilization.This could be due to the fact that moderate N application promoted crop root growth, soil microbial activity, and the decomposition of soil organic matter, and consequently increased soil CO 2 emissions [50].Another reason is that the decomposition of urea in soil by urease can directly generate CO 2 , further increasing CO 2 emissions from urea application treatments [51,52].However, the CO 2 emission factor from urea for warm and cold cropping seasons in soils in the North Plain of China (the research area) was still unclear, which need to be further explored.
The lowest CO 2 emissions were observed under the DCRU treatment, except for CK treatment.This is consistent with previous studies reporting that reduced N fertilizer and CRU application had a positive effect on CO 2 reduction [52,53].DCD application partially inhibited soil CO 2 emission [44].The results of a 3-year field study showed that the application of DCD decreased CO 2 emissions by 7% [25].Another study reported that DCD minimized CO 2 emissions from acidic soils [54].Similarly, Raza et al. [26] also stated that DCD significantly decreased CO 2 emissions from calcareous soils.However, few studies have reported the mechanism of DCD effects on soil respiration and mineralization, which needs to be varied further under field conditions.
Soil CO 2 is mainly produced by crop root respiration and decomposition of organic matter by soil microorganisms, which is affected by soil temperature and moisture con-tent [55].Our results showed slightly higher CO 2 emissions in spring than in autumn, and this can be associated with higher soil temperature and stronger microbial activity in spring than in autumn [56].The highest CO 2 fluxes were observed in the seedling stage, except those of CRU and DCRU at the lotus stage in spring.This may have resulted from the altered soil nutrient availability and carbon supply under N fertilizer application [57].After basal fertilization, the chemical N fertilizer application increased the contents of inorganic N in soil, and provided a sufficient N source for microorganisms, thereby increasing the soil CO 2 emissions in the seedling period.At the lotus stage in spring, CRU and DCRU released more N due to the higher soil temperature, and provided sufficient N for soil microbial activities, which could be attributed to increased soil CO 2 .This study observed that the CO 2 fluxes from the seedling stage in autumn were higher than those in spring.A possible reason may be that, in the seedling stage, the higher temperature and WFPS of the soil promote microbial respiration and decomposition activity, and increased the rate of CO 2 produced by urea decomposition in the autumn over that in spring.

Responses of Lettuce Yields and GHGI to N Application
Within a certain range of N fertilizer application rates, the crop yields and quality increased with increasing N inputs [58].Nevertheless, excessive N application could not improve yield, and caused low N fertilizer utilization efficiency, which not only wastes resources but also reduces soil quality [59].These results were supported by our study, showing that neither reducing N inputs nor applying CRU fertilizer led to a reduction in lettuce yield.The previous studies confirmed that CRU had the potential to balance wheat yield and NUE by promoting plant N uptake and utilization due to the rational supply of nutrients, without decreasing grain productivity compared with conventional urea, even in soils with high fertility [60,61].Similar to previous studies, our results also showed that CRU simultaneously improved lettuce yield and net yield compared with conventional urea at the same N rate.In contrast to CRU, urea dissolved rapidly into the soil after fertilization, resulting in the loss of a large amount of N that failed to be absorbed by plants in time, and making it difficult for plants to absorb sufficient N for growth at the middle and late growth stages, thus decreasing lettuce yield and N use efficiency [34].The CRU may synchronize N release with crop demand compared to traditional fertilizers [4] and can better meet the crop nutrient requirements and improve plant N uptake (Figure 6), contributing greatly to lettuce N accumulation and yield [61]. Notably, our results indicated that the net yield could be significantly increased under long-term use of DCRU fertilizers by 5.41-7.71%,due to the simultaneous limitation of N fertilizer release and inhibition of nitrification in soil.Therefore, DCRU application is a simple and efficient N management strategy to obtain target vegetable yields and economic benefits while increasing NUE and alleviating environmental risks.However, the effect of DCRU on the NUE and yields with changes in soil inorganic N content were unclear, and it is worth exploring in the following experiment.
The mean GWP decreased with N application rate, but the inhibitory effect under ON was much less than that of CRU and DCRU (Figure 4).In comparison with conventional urea, DCRU significantly decreased the GWP.The decrease in GWP under CRU and DCRU was mainly due to the reduced N 2 O emissions, which was in accordance with previous studies [16,18,27].Therefore, the reasonable application of DCRU may be an important measure for slowing down GHG emissions from vegetable systems in the future.
The GHGI is a comprehensive indicator of the greenhouse effect and the economic benefits of farmland [31,62,63].We found that compared with conventional urea, CRU and DCRU decreased GHGI.Meanwhile, GHGI was found to decrease by CRU in the previous study [10].Therefore, the trade-offs between crop yields and GHG emissions in vegetable fields could be obtained by controlled-release fertilizer combined with DCD.
This study provided a scientific basis for promoting environmentally friendly, lowcost, and efficient fertilization practices to achieve global C neutrality.However, to date, our understanding of the effects of the fertilizer coating materials on soil and plants and the long-term effects of different fertilization methods is still insufficient.Future research should take various soil properties and agricultural practices into consideration, and investigate the long-term effects of controlled-release N fertilizer combined with DCD on soil and crops.

Study Site
A five-year experiment was conducted from 2017 to 2021 at Yongsheng Garden agricultural planting center (116 • 41 ′ 32 ′′ E, 39 • 41 ′ 2 ′′ N) located in Tongzhou District in the suburb of Beijing in northern China.It has a temperate continental monsoon climate, with an average annual air temperature of 11.3 • C and an average annual precipitation of 620 mm (mostly occurring in July and August).The soil at the experimental site is classified as sandy soil (sand 55.8%, clay 34.9%, silt 9.3%) according to the Chinese Soil Taxonomy System.A 6 yr wheat-maize rotation was carried out prior to this experiment.The soil characteristics (0-20 cm) are listed as follows: pH 8.12, organic carbon 1.58 g kg −1 , NH 4 + -N 1.53 mg kg −1 , NO 3 − -N 9.80 mg kg −1 , total nitrogen 1.90 g kg −1 .For the U and ON treatments, 40%, 20%, 20%, and 20% of the total N was applied as basal, first, second, and third topdressing fertilizers at the seedling, lotus, and heading stages, respectively.For the CRU and DCRU treatments, the total N was applied all at once as basal fertilization.For all treatments, P and K fertilizers were applied at rates of 78 kg P ha −1 and 184 kg K ha −1 , and all the P and K fertilizers were applied as basal fertilization.Basal fertilizer was surface broadcast by hand, whereas the topdressing fertilizers were dissolved and applied with drip irrigation.The irrigation amount during the entire growth period was 945 m 3 ha −1 .The U and ON treatments had topdressing irrigation every 10 days, a total of 3 times, with each fertilization irrigation of 120-150 m 3 ha −1 , and the CK, CRU, and DCRU treatments were irrigated with equal amounts of clean water.The experiment was repeated for five years (from 2017 to 2021).

Sampling and Measurement
Six fresh plants were randomly collected from each plot at the mature stage, and the fresh lettuce plants were weighed for lettuce yield, then removed the inedible parts and weighed to obtain the net lettuce yield.All lettuce plants were dried in an oven at 70 • C for over 48 h and then weighed for the aboveground dry matter yield.The N content of aboveground lettuce biomass was determined using the Kjeldahl digestion method.
Six soil samples (0-20 cm) were randomly collected from each plot each time at the seedling (on the 6th day after basal fertilizer), rosette (on the 5th day after first topdressing), heading (on the 5th day after third topdressing), and mature stages during the growing season of the lettuce using a stainless-steel soil sampling auger (diameter 4.5 cm).The collected soil samples in the same plot were homogeneously mixed into a composite sample for further chemical analysis.The CRU particles were picked out to ensure that the soil does not contain fertilizer particles.
The soil pH was measured by a potentiometer (Delite B1020, Beijing, China) using a water-to-soil ratio of 5:1.Soil organic carbon (SOC) and total N was determined using an elemental analyzer (Flash Smart NC SOIL, Beijing, China).Soil NO 3 − -N and NH 4 + -N were extracted by dissolving 20 g of fresh soil with 100 mL of 1 mol L −1 KCl solution.The soil extracts were then colorimetrically detected for NO 3 − -N and NH 4 + -N contents using a continuous flow injection analyzer (Auto Analytic 3, Seal Analytical, Germany).
The soil emissions of N 2 O and CO 2 were measured by the static closed-chamber technique in the seedling, rosette, heading, and mature stages.The chamber consists of a plexiglass base collar (50 cm L × 40 cm W × 20 cm H) and a removable top chamber (50 cm L × 40 cm W × 50 cm H) equipped with a battery driven 12 V fan at the center of its inner top.The gas measurements were performed between the lettuce plants.Throughout the lettuce growing season in each year, the sampling was carried out in the morning between 9:00 am and 11:00 am every 2 days for 8 days during the base fertilizer period, every three days after every topdressing N fertilization or irrigation event for 10 days, or once every 7 days otherwise.Four gas samples were successively collected per plot at an interval of 10 min (0, 10, 20, and 30 min) from the chamber headspace using a 25-mL gas-tight syringe, then immediately transferred to 12 mL air-evacuated gas-tight glass vials.The samples were analyzed by using a gas chromatograph (Agilent, HP7890, Agilent, Santa Clara, USA) equipped with an electron capture detector (ECD) to detect N 2 O and a flame ionization detector (FID) to detect CO 2 [64].The CO 2 and N 2 O exchange fluxes were calculated using Equation (1) [65]: where F in CO 2 is mg m −2 h −1 and N 2 O is µg m −2 h −1 , ρ is the concentration of N 2 O and CO 2 (e.g., N 2 O: 1.977 kg m −3 , CO 2 : 1.997 kg m −3 ).h is the height of the effective space in the chamber (m), ∆c is the gas concentration difference, ∆t is the time interval, and T is the mean soil temperature for each sampling ( • C).
The cumulative emissions of CO 2 and N 2 O were calculated for each treatment according to Equation (2) [31]: where CGE is the cumulative emission (kg ha −1 ) for each gas, n is the total sampling time, V i and V i+1 are the measured fluxes of two consecutive sampling days, and (t i+1 − t i ) is the time interval between two adjacent sampling days.The warming potentials (GWP) of N 2 O are 298 times that of CO 2 , on a 100-year timescale [2]; therefore, the equation for calculating the GWP (kg CO 2 -eq ha −1 ) is Equation (3): where GCO 2 (kg CO 2 ha −1 ) and GN 2 O (kg N 2 O ha −1 ) are the cumulative CO 2 and N 2 O emissions, respectively.Greenhouse gas intensity (GHGI) refers to the comprehensive warming potential per unit of output, and is calculated according to Equation (4): where GHGI is the GHG emission intensity (kg CO 2 -eq kg −1 lettuce yield), GWP is the total amount of greenhouse gas emissions (kg CO 2 eq ha −1 ), Y is the yield of lettuce (kg ha −1 ).

Statistical Analysis
Statistical analysis was performed using SPSS 24.0 (IBM Co., New York, NY, USA).Before the analysis, a Shapiro-Wilk test and Levene test were employed to check the normality and variance homogeneity of the data.One-way analysis of variance (ANOVA) was performed to assess the effects of different treatments on lettuce yield, N uptake, soil NH 4 + -N and NO 3 − -N content, and N 2 O emissions.The means of significant effects were compared using Duncan's multiple range test (p < 0.05).All the figures were drawn with Origin 2023 (OriginLab Co., Northampton, MA, USA).

Conclusions
In comparison with conventional urea, N 2 O emissions reduced but the reduction effect on CO 2 emissions was not obvious, and lettuce yield and net yield were maintained in ON.CRU was effective in controlling N 2 O by prolonged NH 4 + -N availability in the soil of the open-field lettuce system and in improving the N content of lettuce and net yield.DCRU was the most effective in mitigating N 2 O and CO 2 fluxes and the mean cumulative effect of N 2 O and CO 2 emissions, with a reduction of up to 37.68% and 9.28%%, respectively, and the net yield was also significantly increased by 5.41-7.71%,due to the improved N absorption of the lettuce in higher soil NH 4 + -N, lower NO 3 − -N content under the long-term use of DCRU.Therefore, the simple and efficient N management strategy to obtain trade-offs between GHG emissions and lettuce yields in the open-field lettuce system could be obtained by controlled release of fertilizer containing DCD without topdressing.This study provided a theoretical basis for the efficient and sustainable development of open-field vegetable production by adopting optimized N management practices.

Figure 2 .
Figure 2. Soil CO2 emission flux during the experimental period (2017-2021) under different treatments.Note: S-seedling, L-lotus, H-heading, M-mature, Spr.-spring, Aut.-autumn.CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.The cumulative emissions of N2O and CO2 were consistent with the patterns of N2O and CO2 fluxes (Figure 3a,b).The cumulative emissions of N2O and CO2 were 3.95-7.51kg ha −1 and 9127.18-11003.84kg ha −1 in the N application treatments, respectively, and decreased under CRU and DCRU treatments in the 2017-2021 period, achieving the

Figure 3 .
Figure 3.The mean cumulative emissions of N 2 O (a) and CO 2 (b), GWP (c) and GHGI (d) during experimental period (2017-2021) under different treatments.Note: Means with the same lower-case letter across treatments within each figure are not significantly different at p < 0.05.The error bars represent the standard error.CK, without N application; U, according to local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.

Figure 4 .
Figure 4. Soil NH4 + -N ((a), spring; (b), autumn) and NO3 − -N ((c), spring; (d), autumn) contents during the experimental period (2017-2021) under different treatments.Note: Means with the same lower-case letter across treatments within each figure are not significantly different at p < 0.05.The error bars represent the standard error.CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.

Figure 5 .
Figure 5.The lettuce yield ((a), spring; (b), autumn) and net yield ((c), spring; (d), autumn) dur the experimental period (2017 to 2021) under different treatments.Note: Means with the sa lower-case letter across treatments within each figure are not significantly different at p < 0.05.T error bars represent the standard error.CK, without N application; U, according to the lo farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated u at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% rate.

Figure 5 .
Figure 5.The lettuce yield ((a), spring; (b), autumn) and net yield ((c), spring; (d), autumn) during the experimental period (2017 to 2021) under different treatments.Note: Means with the same lower-case letter across treatments within each figure are not significantly different at p < 0.05.The error bars represent the standard error.CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.

(g kg − 1 )Figure 6 .
Figure 6.Lettuce N contents in spring (a) and autumn (b) during the experimental period (2017-2021) under different treatments.Note: CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.

Figure 6 .
Figure 6.Lettuce N contents in spring (a) and autumn (b) during the experimental period (2017-2021) under different treatments.Note: CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.
Four different N fertilizer management treatments were established: (1) CK: without N fertilizer; (2) U: conventional urea fertilizer (46% N, according to local farmers' practice) applied at 300 kg N ha −1 ; (3) ON: conventional urea fertilizer applied at 240 kg N ha −1 ; (4) CRU: polyurethane-coated urea (42% N, 60-day release, developed by the Institute of Plant Nutrition, Resources and Environment at the Beijing Academy of Agriculture and Forestry Sciences) at 240 kg N ha −1 ; (5) DCRU: polyurethane-coated urea containing DCD (DCD mixed with urea was coated by polyurethane after granulation, DCD:N = 1:100) at 240 kg N ha −1 .All treatments were arranged in a randomized complete block design with four replicates.The area of each plot was 28 m 2 (7 m × 4 m).A ridge with 0.3 m width and 0.3 m height was set up to prevent the exchange of water and nutrients between plots.The open field was planted with the lettuce (Lactuca sativa) 'Sheshou No. 101', which is a common local cultivar with a growth period of approximately 60 days in the spring and autumn.The lettuce seedlings were transplanted on 15 April or 18 April and 19 August or 23 August in the spring and autumn, respectively.

Table 2 .
Effect of different treatments on CO2 emission fluxes (mg m −2 h −1 ) in the lettuce growing seasons.
Figure 1.Soil N 2 O emission flux during the experimental period (2017 to 2021) under different treatments.(Note:S-seedling, L-lotus, H-heading, M-mature, Spr.-spring, Aut.-autumn).CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethanecoated urea containing DCD at a reduced 20% N rate.Note: Values are means ± SE (standard error) of four replicates.The same letter within row means no significant difference at the level of 0.05.CK, without N application; U, according to the local farmers' practice; ON, conventional urea at a reduced 20% N rate; CRU, polyurethane-coated urea at a reduced 20% N rate; DCRU, polyurethane-coated urea containing DCD at a reduced 20% N rate.

Table 2 .
Effect of different treatments on CO 2 emission fluxes (mg m −2 h −1 ) in the lettuce growing seasons.