Reactive Nitrogen Releases And Nitrogen Footprint During The Life-Cycles of Intensive Vegetable Production Affected By Human Feces Slurry Substitution

Evaluating the sustainability of vegetable production is crucial to secure future food supply. A two-year eld study of four different vegetable crops was performed to investigate the effects of inorganic fertilizer and human feces slurry at different ratios on vegetable yields, reactive gaseous nitrogen emissions (GNrEs), reactive nitrogen (Nr) footprint and net ecosystem-economic income (NEEI) by using life-cycle analysis. Four fertilization strategies were studied, including: CK (no fertilization); CF (inorganic fertilization); CHF1 (human feces slurry/inorganic fertilizer, N ratio=1:7); and CHF2 (human slurry/inorganic fertilizer, N ratio=1:3). Results showed that compared with CF treatment, both CHF1 and CHF2 treatments increased the N 2 O+NO emissions by 11.8 % and 32.4 % on average, while decreased the vegetable yields by 6.7 % and 7.4 %, respectively. Moreover, the addition of human feces slurry increased the proportions of Nr footprint by 6.6 % (CHF1) and 2.9 % (CHF2) in comparison with CF treatment group. However, although CHF2 treatment signicantly increased the values of GNrEs and reactive gaseous nitrogen intensity (GNrI) by 8.4 % and 12.5 %, respectively, in relation to those in CF treatment group, it still increased farmers’ income by 16,404 CNY ha −1 . These ndings suggest that although human feces slurry incorporation could not mitigate Nr releases, the appropriate ratio of inorganic fertilizer and human feces slurry (CHF2) is able to improve net economic income (NEI) and NEEI during intensive vegetable production. Nevertheless, the relationship between combinatorial treatment of inorganic fertilizer and human feces slurry and mitigation of Nr release should be explored further. Government situ of feces resource. It foreseeable that the rapid development of rural toilets, the human feces slurry and FID). The concentrations of NO were measured by a Thermo model 42i chemiluminescence NO–NO–NO X analyzer (Thermo Environmental Instruments Inc., USA). NH 3 volatilization uxes were assessed with a continuous air-ow enclosure method (Sun et al., 2017), and the measurement device consisted of a chamber, two acks a vacuum pump and a vent pipe. The absorbent for NH 3 was 80 mL of 2% w/v boric acid, and the absorption was determined by titration with 0.01 M H 2 SO 4 using a mixture of bromocresol green and methyl red (in ethanol solution) as an indicator. NH 3 volatilization was measured once a day after fertilization until no volatilized gas was detectable (about 10 days). The gaseous reactive nitrogen emissions (GNrEs, kg·N·ha − 1 ) and the reactive gaseous N intensity (GNrI; kg·N·t − 1 yield) was calculated using the following equations:


Introduction
Since 2004, Chinese government has invested RMB 8.38 billion in the construction and renovation of 21.263 million rural toilets, and the coverage of sanitary toilets in rural areas has increased from 7.5% in 1993 to 78.5% in 2015 (NHFPC, 2016; Cheng et al. 2018). Great achievements have been made in the rural toilet renovation, but the inevitable consequence is a large amount of feces slurry with low reutilization e ciency (Koger et al. 2014). Human feces treated in septic tanks can be used as fast-release fertilizer with considerable amounts of nitrogen, phosphorus and potassium compounds, and the Government has encouraged the in situ utilization of feces resource. It is foreseeable that with the rapid development of rural toilets, the proportion of human feces slurry may be increased further.
Vegetable is the second largest crop in terms of planting area (after grain) in China . The annual yield of Chinese vegetables was 76.9 million tons, and the sowing area reached 2.2 million ha, accounting for 13.2% of the global sown area . In the past, a relatively high economic value of growing vegetables has encouraged farmers to use fertilizers for maximizing crop yields , which may lead to serious environmental problems (Kim et al. 2006). But now, more and more farmers have realized that the sustainable utilization of natural resources, such as human feces slurry, could decrease the use of inorganic fertilizer as well as reduce the production costs (Zhou et al. 2019).
Nowadays, it is well known that intensive vegetable production has become an important source of reactive nitrogen (Nr) releases in China, owing to the large application amounts of N fertilizer and frequent irrigation events (Fan et al. 2017). Approximately 20-50 % of applied fertilizer-N is lost as Nr, such as gaseous emissions of ammonia (NH 3  However, different substitute resources and different incorporation ratios might potentially result in variable effects on Nr releases and crop productivity due to diverse fertilizer types and N transformation patters in agricultural systems (Zhang et al. 2012). More importantly, there are only limited data available on the effects of combination treatment of human feces slurry and inorganic fertilizers at different ratios on Nr losses and crop productivity during intensive vegetable production.
Evaluating the magnitude of the impact of N loss on agroecosystems can generate possible solutions to mitigate climate change and other environmental problems, thus helping to raise awareness in the general public and facilitating decision making with respect to environment-friendly technological development by policy makers. In recent decades, nitrogen footprint has been proposed as a potential indicator to assess how individuals, communities, organizations, or countries contribute to nitrogen pollution through their consumption, and thereby affect the environment and human health. It is widely known as the "total amount of Nr [reactive nitrogen, all other forms than N 2 ] released to the environment as a result of N consumption" (Leach et al. 2012). In fact, nitrogen footprint has been identi ed as a critical member of the "footprint family", i.e., a more comprehensive measure of of human impact on agroecosystem (Galli et al. 2012;Leach et al. 2012).
In the present study, a two-year eld experiment was performed on four different vegetable crops to quantify the amounts of Nr, such as N 2 O, NO and NH 3 emissions, after treatment with mineral N fertilizer and human feces slurry in an intensively managed vegetable eld in the North China. The aim of this research was to evaluate the impact of different portions of inorganic fertilizer and human feces slurry on average Nr footprins and net ecosystemeconomic income (NEEI) associated with the loss of N during intensive vegetable production. We hypothesized that the combination treatment of inorganic fertilizer and human feces slurry at different ratios could affect the proportions of Nr footprint and NEEI during intensive vegetable production.
According to the Tianjin weather station, this area was dominated by subtropical monsoon climate with an average annual rainfall of 642.8 mm and a mean annual air temperature of 11.2°C. The studied soil was classi ed as Cambosol (equivalent to Inceptisol in the USDA Soil Taxonomy), and its main compositions included total N: 1.2 g·kg − 1 , soil organic carbon (SOC): 9.7 g·C·kg − 1 , cation exchange capacity (CEC): 16.3 cmol·kg − 1 and pH: 8.4. Human feces slurry were collected from a household septic tank nearby the experiment site, after 60 days of anaerobic fermentation, the human feces slurry can basically meet the requirements of harmlessness, and the details of the human feces slurry are listed in Table 1.

Experimental treatments and vegetable management
The experiment was carried out in 3 replicate plots arranged in a randomized complete block design (each plot: 3 × 2.5 m). There were four treatment groups as follows: (i) CK, no fertilization; (ii) CF, inorganic fertilization; (iii) CHF1, combination of inorganic fertilizer and human feces slurry with a N ratio of 7:1; and (iv) CHF2, combination of inorganic fertilizer and human feces slurry with a N ratio of 3:1. In addition, four different vegetable crops were successfully grown in each year, such as fennel (Foeniculum vulgare Mill.), Chinese cabbage (Brassica rapa L.), spinach (Spinacia oleracea L.) and lettuce (Lactuca sativa L.), and there was a short fallow after the harvest of each crop. The same vegetable rotation was performed in the second year, indicating that a total of eight vegetable crops were harvested during this eld experiment. The fertilizer rates of urea (N = 46.4%), calcium superphosphate (P 2 O 5 = 60.7%), potassium chloride (K 2 O = 63.1%) and human feces are listed in Table 2. All the fertilizers were incorporated into the soil 3-4 days before sowing. The rates of N application were set according to the local vegetable cropping regimes and fertilizing practices. Meanwhile, the crops in CK group followed the eld management practices as similar to those in N-supplied treatment groups. All vegetable elds were plowed before transplanting or sowing. According to the local practices, the application of N fertilizer was often paired with irrigation. Table 2 Annual fertilizer application rates for the different experimental treatments of vegetables (kg ha − 1 ).  The gaseous reactive nitrogen emissions (GNrEs, kg·N·ha − 1 ) and the reactive gaseous N intensity (GNrI; kg·N·t − 1 yield) was calculated using the following equations: GNrI = GNrEs/vegetable yield(fresh yield) 2.4 Vegetable yield, above-ground N uptake and use e ciency After reaching physiological maturity, the above-ground parts of each vegetable crop in each plot were weighted and recorded as fresh weight.
Subsequently, total N uptake was calculated from the biomass of harvested crop within each plot. The crop biomass was air-dried, and then oven-dried at 65°C for 3 days. After drying, the yield of dry matter was determined. Subsamples were then ground using a ball mill, and their N content was evaluated by a FOSS N analyzer (KT260, Foss Co., Germany). Nitrogen use e ciency (NUE) was calculated using the following equation: where TU N−CK is the difference in the N content of the above-ground crops between N-supplied treatment groups and control group (kg·ha − 1 ); and TN is the total input of fertilizer N (kg·ha − 1 ).

Estimation of Nr footprint, ecosystem input and ecosystem-economic income
The Nr footprint (g·N·kg − 1 food) was calculated as follows: where AI iNr indicates the loss of Nr (primarily through NO X and N 2 O emissions) during the harvesting and postharvest handling of agricultural inputs, and the values are shown in Table S1; FC jNr denotes the loss of Nr during farm cultivation (primarily through NO X and N 2 O emissions, N leaching and runoff, and NH 3 volatilization). The amounts of N 2 O and NO emissions were estimated in this eld study by using both static chamber and gas chromatography methods, and other Nr species were determined by multiplying the amount of fertilizer by their individual emission factor as described by Ti et al. (2015), i.e., 0.14 and 0.12 for open-air and greenhouse vegetable cropping systems, respectively. The vegetable yields used in this formula were dry matter yields.
The environmental external cost is constituted of the global warming associated with greenhouse gas emissions, soil acidi cation linked to NH 3 and NO x emissions, and aquatic eutrophication resulted from NH 3 emission as well as N leaching + runoff (Xia and Yan, 2012). Here, the ecosystem input (CNY ha − 1 ) associated with N loss was assessed using the following equation: where Nr i A (kg N) indicates the released amount of speci c Nr forms; and Pi (¥ kg − 1 N) is the cost to climate change, ecosystems and human health per kg of speci c Nr (Table S2).
Net economic income (NEI; CNY ha − 1 ) and net ecosystem-economic income (NEEI, CNY ha − 1 ) for crop production were determined using the following equations: where 'yield income' (CNY kg − 1 ) represents the market value of each crop; 'input cost' (CNY kg − 1 ) denotes the cost (per kg) incurred during food production, including procurement of labor and agricultural materials; and 'environmental external cost' (CNY kg − 1 ) indicates the cost (per kg) incurred during food production as a consequence of damage triggered by Nr release.

Statistical analysis
Statistical analysis was carried out using JMP version 9.0 (SAS Institute Inc., Cary, NC, USA, 2010). All data were normally distributed and had homogeneous variances, hence, the experimental results were analyzed with a parametric test. The statistical comparison the annual and seasonal cumulative emissions of N 2 O and NO, vegetable yield, and Nr footprints among different treatment groups was conducted using an one-way analysis of variance (ANOVA). Tukey's multiple range test was employed to assess whether there is any signi cant difference between the treatment groups at a signi cance level of < 0.05.

Reactive gaseous nitrogen emissions
During the vegetable production period, N 2 O and NO emissions peaked after the application of N fertilizer, tailed off after one week, and then remained low (Fig. 1a). Notably, the peaks of NO uxes were lower than those of N 2 O (Fig. 1b). The seasonal N 2 O emissions occurred mainly in the fennel planting season, which might attributed to the relatively high temperature. The largest NO ux peak occurred in the Chinese cabbage growing season after treatment with CHF2. Despite relatively lower temperature during the planting seasons of spinach and Chinese cabbage, several NO uxes peaks were still observed, mainly due to a decrease soil moisture that is conducive to NO emissions.
The cumulative reactive gaseous nitrogen emissions generated from different treatments are presented in Table 3. The average seasonal cumulative N 2 O, NO and NH 3 uxes of the four treatment groups ranged from 5.9-17.1 kg·N·ha − 1 , 1.6-5.4 kg·N·ha − 1 and 5.1-12.7 kg·N·ha − 1 , respectively. Notably, CHF2 treatment signi cantly increased N 2 O emissions by 44.9 % compared to the CF treatment groups, respectively (p < 0.05; Table 3). The lowest average cumulative NO emissions were found in CHF1 group among the three N-supplied treatment groups, but the differences were not statistically signi cant (p > 0.05). The lowest NH 3 emission was recorded in CHF2, which was remarkably lower than that in CF (p < 0.05). In addition, CHF2 treatment signi cantly increased GNrEs and GNrI by 8.4% and 12.5% when compared to CF treatment group (p < 0.05), and these values were the highest among the three N-supplied treatment groups.

Vegetable yield and economic bene ts
The vegetable fresh yield and economic bene ts among all the treatment groups are shown in Table 4. Not unexpectedly, it was found that the three Nsupplied treatment groups (CF, CHF1 and CHF2) markedly elevated the yields of vegetable (91.3%, 77.2% and 78.5%, respectively) compared to CK treatment group (p < 0.05; Table 4). The highest average yield (190.2 t·ha − 1 ·yr − 1 ) of vegetable was observed in CF treatment, and as a comparison, the two combinatorial treatments with human feces and inorganic fertilizer (CHF1 and CHF2) decreased the average vegetable yields by 7.4% and 6.7%, respectively. However, no signi cant difference in vegetable yields was noted among all the N-supplied treatment groups. The others contained plastic lm and pesticide. Table 4 also shows the total economical balance for the input resources, the output pro ts, and the environmental cost for different treatment groups averaged over two years. The inputs for agricultural production, such as seeds (34.6-39.4%) and fertilizer (46.4-55.3%), constituted an important part of the total input of the N-supplied vegetable cropping system. Given that most farmers have enough family laborers to work on vegetable production, the labor cost generally included septic tank cleaning only. In addition, compared to CF treatment, CFH2 and CFH1 treatments increased (3.7 %) and decreased (4.3 %) the farmer's economic income, respectively. The environmental external cost associated with the GNrE as well as the leaching and runoff caused by N application should be considered due to their in uences on the society. The total environmental external cost varied from 732 to Page 7/13

N uptake and NUE
Higher N uptake (0.69 kg·N·ha − 1 ) was observed in CF treatment group, whereas the addition of human feces could decrease N uptake compared to CF treatment group, but the differences were not statistically signi cant (p > 0.05; Fig. 2a). Taking the amounts of N uptake into considerations, we also quanti ed the NUE in this study (Fig. 2b). The NUE of all the four vegetable crops ranged from 16.8-20.3% among all the N-supplied treatment groups.
In agreement with the results of vegetable yields, CHF1 and CHF2 treatments both decreased the NUE values by 17.2% and 5.4%, respectively, when compared to CF treatment group, but statistical signi cance was only found in CHF1 treatment group (p < 0.05).

Nr losses and Nr footprint
The estimated total losses of Nr for all the N-supplied treatments varied from 177.6 to 184. The Nr footprint of vegetable production varied from 1.7 (CK) to 11.7 g·N·kg − 1 (CHF1). Compared to CK treatment, N application remarkably increased the value of Nr footprint (p < 0.05; Fig. 3). In addition, a consistent effect of the combination of inorganic fertilizer and human feces slurry at different ratios was detected in both CHF1 and CHF2 treatment groups. Compared to CF treatment, the proportions of Nr footprint were increased by 4.4 % in CHF1 group and decreased by 0.5 % in CHF2 group (p > 0.05).

Effects of different fertilization strategies on GNrEs
The N fertilizer used in the eld could increase the initial levels of NH 4 + and NO 3 − during nitri cation and denitri cation processes (Cardoso et  Generally, greenhouse gas intensity and global warming potential (GWP) are used to integrate the effects of greenhouse gases. In our research, GNrEs and GNrI were employed to investigate the overall impact of gaseous reactive nitrogen on the environment. Table 3 showed that N 2 O and NH 3  We estimated the NUE in all the N-supplied treatments, and found that the NUE ranged from 16.8-20.3 %, which were higher than the values reported by Li et al (2017). However, the two treatments incorporating human feces slurry decreased the NUE in relation to CF treatment (Fig. 2), indicating that replacing the inorganic fertilizer with human feces slurry could not effectively increase the NUE during intensive vegetable production. This could be attributed to the higher Nr emissions in CHF1 and CHF2 groups than those in CF group. Generally, substituting compound fertilizers (usually by organic matter) markedly elevated plant N uptake and reduced gaseous N losses Zhuang et al. 2019), however, the composition of human feces slurry was different from that of standard organic fertilizer, despite containing a certain amount of organic matter, and human feces usually considering as quick-acting fertilizer rather than organic fertilizer.
In this study, human feces replacing part of nitrogen fertilizer exerted no pronounced effects on the vegetable yield in two years (four successive vegetable seasons), possibly due to the fact that the total N application rates were relatively similar in the three N-supplied treatment groups (Table 4). It has been shown that with NH 4 + as the dominant nitrogen form, both grain yields and N uptake improved than the conventional fertilizer strategy (Deppe et al. 2016). However, in this study, the vegetable yields were similar between CHF1 and CHF2 treatment groups, but did not differed greatly from CF treatment group. This was probably due to the fact that we applied human feces slurry as base fertilizer before planting/seeding, suggesting that the high NH 4 + concentration immediately upon application could not induce toxicity (cf. Müller et al., 2006). Indeed, the results of crop yields were in good agreement with the enhanced uptake of nutrients during the growing season (Cardenas et al. 2016).
Emissions of NO x , N 2 O and NH 3 can threaten human health and cause severe diseases (e.g., cataract, skin cancer, etc.), mainly via particle pollution, ground-level ozone pollution, and stratospheric ozone depletion. In China, the total health damage cost related to atmospheric Nr emissions was estimated at US$19 − 62 billion in 2008, which is much larger than the damage costs in the United States and Europe. According to Gu et al. (2012), agricultural Nr emission is one of the largest sources of health-related damage, accounting for about 50% of total expenditures in China, which are consistent with the results of this study (Fig. 3). However, other reports identi ed NO x emissions as the largest source of health-related damage in the United States and Europe (Birch et al. 2015), which is contrary to our ndings (Fig. 3), possibly because of the high N inputs during intensive vegetable production ( Table 2). Moreover, leaching + runoff was an important contribution to the ecosystem N input in this study, which emphasized the importance of reducing leaching + runoff during intensive vegetable production. In addition, we used the coe cient estimated by Ti et al (2015) related to the total N applied, which might have underestimated the leaching + runoff losses in the two treatment groups with human feces slurry due to its rapid in ltration.
Pro tability is often regarded as the main driver for farmers to improve their agricultural practices. Taking the yield income, input costs and environmental cost into consideration, we also quanti ed the net ecosystem economic bene ts among the three N-supplied treatment groups (Table 4).
Notably, CHF2 had the highest values of NEI and NEEI, representing a su cient incentive for farmers to alter their N management strategies, if such information were available for them.

Nr footprint in crop production
The evaluation of Nr footprint can aid better detection of the 'authentic' environmental hotspots of Nr releases in the national food system (Cheng et al. 2014; Chen et al. 2020). In our research, the Nr footprint ranged from 1.6-11.8 g·N·kg − 1 (Fig. 3), which were comparable to the ndings in a wheatmaize system in North China Plain ) and different fertilization strategies in vegetable crop rotations in Southeast China (Zhou et al. 2019). In addition, our results indicated that the N leaching and runoff, with a high N application rate, was the dominant contribution to the eld Nr footprint, accounting for about 72% of the total Nr footprint. It is commonly known that N fertilizer is an essential source of Nr emissions during crop cultivation (Chen et al., 2014). However, this is inconsistent with Xue and co-workers (2016), who analyzed the Nr footprint of double rice production in Southern China by using the life cycle assessment method, and identi ed that NH 3 emission from paddy elds could be the main contributor. Typically, the Nr footprint in paddy elds is dominated by NH 3 volatilization, whereas N leaching/runoff is a primary contributor to Nr footprint in the upland ecosystems (Xia et al. 2016). It is worth noting that the amount of N leaching and runoff in our study was estimated by the coe cient obtained from

Conclusion
A major obstacle for solving N-driven environmental issues is the lack of knowledge on complete N budgets, such as the major N uxes, in agricultural systems, especially those under alternative farm management practices. Our results indicate that the release of Nr is dominated by N leaching and runoff. The human feces slurry substituting for a portion of inorganic fertilizer elevated Nr footprint and environmental cost when compared to CF treatment. However, although with increasing environmental costs, CHF2 treatment increased farmers' net economic income compared to CF treatment.
Our study suggests that the application of human feces slurry together with inorganic fertilizer at an appropriate ratio could increase farmers' net income, even though releases more Nr into environment. Therefore, further investigations are required to ensure Nr mitigation in cases of human feces slurry treatment during intensive vegetable production.   Table 1 for treatment codes.

Figure 3
Contributions of different sources/activities to the Nr footprints of intensive vegetable production. N fertilizer refers to synthetic N fertilizer production, transportation and application; Others refers to the sum of other sources of GHG, such as the production of phosphorus, potassium and their transportation and application. The small letters in each sub gure indicate signi cant differences according to the Tukey's multiple range test (p<0.05) among all the treatments.