Soil profile N2O efflux from a cotton field in arid Northwestern China in response to irrigation and nitrogen management

It remains uncertain how different N inputs as synthetic fertilizer or manure and irrigation types affect nitrous oxide (N2O) production and effluxes in the subsurface. A field trial was carried out in 2016 to evaluate the impacts of conventional urea, animal manure, and a 50/50 mix of urea and manure on N2O production/effluxes from a cotton (Gossypium hirsutum L.) field under flood or drip irrigation in northwestern China. Soil N2O concentrations were monitored at 5, 15, 30, and 60 cm depths to assess the production and diffusion rates of N2O in the soil profile. The results showed that N2O concentrations in 0–60 cm ranged between 221 and 532 nL L−1 and averaged 344 nL L−1, which was generally lower compared to other studies in the same region. Manure and flood irrigation significantly increased N2O production at 0–5 cm and 5–15 cm, respectively. That is, the effects of nitrogen management and irrigation types on the N2O production of the profile were reflected in the surface layers and subsurface layers, respectively. All N2O production occurred in the 0–15 cm layer, with the 0–5 cm depth contributing 87%–100% of the surface emissions. The response discrepancy of N2O production/diffusion to irrigation and nitrogen management in different soil depths should be fully considered in developing agricultural N2O emission reduction measures.


Introduction
Nitrous oxide (N 2 O) is a powerful greenhouse gas that contributes to both global warming and stratospheric ozone destruction (Ravishankara et al., 2009). Soil N 2 O emissions have rapidly increased from 6.3 Tg N 2 O-N yr −1 in pre-industrial times to 10.0 Tg N 2 O-N yr −1 in recent years, with 82% of the total increase coming from cropland (Tian et al., 2019). Application of manure and synthetic fertilizer is the main factor inducing N 2 O emissions from agricultural soils (Tian et al., 2020). In addition, irrigation practice is also a crucial factor in determining N 2 O emissions from agricultural ecosystems (Kuang et al., 2021).
Thus, it is essential to explore the effects of fertilization and irrigation methods on the production and emission of N 2 O. Irrigation management practices affect N 2 O emissions through their impacts on the spatial and temporal distribution of soil moisture content, as well as microbial and nutrient availability (Sánchez-Martín et al., 2008;Kuang et al., 2018). Drip irrigation is an effective practice to enhance N and water use efficiency, and it is widely used in arid regions for crop production (Vázquez et al., 2006). Some studies reported that drip irrigation effectively reduced N 2 O emissions from cropland, compared with traditional irrigation (Sánchez-Martín et al., 2008;Bronson et al., 2018;Li et al., 2014), whereas in others the opposite was observed (Fentabil et al., 2016;Kuang et al., 2018). A global meta-analysis showed that drip irrigation significantly reduced N 2 O emissions from cropland by 32% and 46% compared to traditional flood and sprinkler irrigation, respectively (Kuang et al., 2021). Using a soil column incubation study, Kuang et al. (2019) revealed that deep-placed N fertilizers were most susceptible to denitrification under high water-filled pore space (WFPS) content but did not result in a significant surface burst of N 2 O emission, suggesting soil moisture plays an essential role in determining production and consumption of N 2 O across soil profiles. Understanding the effects of different management measures on soil N 2 O production processes can provide a basis for optimizing agricultural management practices.
N fertilizer and manure additions are the main cause of N 2 O emissions from agricultural land (Tian et al., 2020). Soil properties, including the form of N and available C can affect the biological processes of nitrification and denitrification (Velthof et al., 2003). The effects of N sources on N 2 O production and consumption have highly complex regulatory mechanisms (Zhou et al., 2017). Several studies have found that manure addition increased N 2 O production compared to synthetic N fertilizers by providing C substrate to denitrifiers for denitrification (Hayakawa et al., 2009;Ju et al., 2011;Forte et al., 2017;Yin et al., 2019). Other studies, on the other hand, have found that manure application reduces N 2 O emissions when compared to synthetic fertilizers, by stimulating complete denitrification to N 2 (Ball et al., 2004;Meijide et al., 2007;Tao et al., 2018). Several studies also found no differences in the use of manure and synthetic N fertilizer in terms of N 2 O emissions (Meng et al., 2005;Vallejo et al., 2006). These inconsistent results reflect the need for further analysis of the effects of different N sources on N 2 O production, transport, and consumption in the profile. As a result, there remains a scarcity of knowledge on the relationship between N 2 O efflux underground and emissions on the surface, which are influenced by N fertilizer and manure with the drip-and floodirrigated crops.
Surface N 2 O emissions are the net result of a series of processes involving profile N 2 O production, diffusion and consumption (Gao et al., 2014;Wang et al., 2018;Li et al., 2021). The rate and direction of N 2 O diffusion are determined by the distribution of N 2 O concentration in the profile. Depending on the concentration gradient of N 2 O in the subsoil, its association with soil surface emission rates can be used to quantify the contribution of N 2 O from different soil layers to surface N 2 O emissions (Nan et al., 2016;Wang et al., 2018). Nan et al. (2016) reported that 99% of the total cumulative N 2 O fluxes in the soil profile occurred in the 0-15 cm soil layer. According to Wang et al. (2018), soil N 2 O consumption at depths of 0-5 and 5-15 cm attributed to 80.4% and 6.6% of the surface N 2 O emission, respectively. However, few studies have been conducted to compare the contribution of different soil layers to surface N 2 O emissions under different irrigation and N management practices.
The objectives of this study were to 1) characterize the spatial distributions of N 2 O concentrations in the soil profile with urea or manure application under drip and flood irrigation, 2) quantify the depth-dependent contributions of profile N 2 O effluxes to the surface emissions, and 3) assess the impact of environmental factors on N 2 O fluxes in the profile.

Site description and soil properties
A field experiment was carried out at the Cele National Station (37°01′06″N, 80°43′48″E) in Xinjiang Uygur Autonomous Region, during the 2016 growing season. The station is situated on the southern edge of the Taklimakan Desert. The mean annual precipitation and annual potential evaporation are 42.5 mm and 2,956 mm, respectively. The average annual air temperature is 12.7°C. The soil is classified as Aridisols in the USDA ST system (USDA, 1999), and the surface soil (0-20 cm) has sand, silt, and clay content of 900, 40, and 60 g kg −1 , respectively. For details about the soil properties as shown in Table 1.

Experimental design and crop management
This experiment was a two-factor experiment in a randomized complete block design with two types of irrigation (drip and flood) and four N source treatments: 1) no fertilization (Control), 2) granular urea (Urea), 3) animal manure (Manure), and 4) 50% granular urea with 50% animal manure (U + M). Detailed information about the experimental design has been described in our previous study (Kuang et al., 2018). Briefly, the application rate of all fertilizer treatments was 240 kg N ha −1 . Granular urea (N 46%) was applied as 522 and 261 kg ha −1 for Urea and U + M, animal manure was applied as 76.9 and 38.5 Mg ha −1 for Manure and U + M, respectively. Under drip irrigation, 20% urea was applied at planting, with the remaining 80% applied as a topdressing six times during the growing season. Under flood irrigation, 30% of the urea was applied during planting, and the remaining 70% was topdressed to the soil four times before irrigation. Under both irrigation systems, manure was evenly broadcast over the surface soil before sowing and immediately incorporated with the soil. For all plots, calcium phosphate (120 kg P 2 O 5 ha −1 ) and K 2 SO 4 (60 kg K 2 O ha −1 ) were broadcast on the surface and mixed into soils (0-20 cm) with a rota-cultivator before planting. Each treatment had four replicated plots. In total, 32 plots were set up in our study, each plot with an area of 32 m 2 (10 m × 6.4 m).
Planting and crop management were described in our previous studies (Kuang et al., 2018;Li et al., 2020). Briefly, plastic film was used to cover four cotton rows with row spacing of 30-50-30 cm. For each plastic film in drip-irrigated treatments, drip tapes were installed between two cotton rows (30 cm apart) and the distance Frontiers in Environmental Science frontiersin.org between every two emitters was 10 cm. The water flow rate in emitter was 2-3 L h −1 . In each plot, water and fertilizer-integrated tanks were placed to record the amount of irrigation and urea application. Over the experimental period, cotton in the drip irrigation system had received 9 times of irrigation, with each irrigation providing approximately 45 mm water. In contrast, cotton in the flood irrigation system had received 7 times irrigation of approximately 140 mm water for each irrigation.

Soil N 2 O gas sampling and analysis 2.3.1 Surface N 2 O emissions
The static chamber method was used to monitor soil surface N 2 O flux (Kuang et al., 2018). The sampling frequency was once or twice per week to make sure a sampling was done within 1-2 days after irrigation and fertilization events. The sampling time was 10: 00-14:00 (GMT+8) during the day, and N 2 O concentration in gas samples during this time period was used to represent the daily average.

Profile N 2 O collection
Soil profile N 2 O concentration was measured simultaneously with the surface N 2 O emissions. Soil profile N 2 O gas at depths of 5, 15, 30, and 60 cm were collected using an in-situ soil profile gas sampler (for more details, see Kuang et al. (2019)). Briefly, the gas sampler was composed of four individual silicone tubes (5.0 cm long, 36.8 mm i. d., 40.0 mm o. d.) that air but not water can go through and sealed at both ends. The silicone tube was covered by a polyethylene (PE) pipe (5.0 cm long, 40.8 mm i. d., 50.0 mm o. d.) to determine the soil N 2 O gas sampling depth by the holes in the wall of the PE pipe. A hollow stainless-steel tube (0.6 mm i. d.) with a sampling port was used to collect soil N 2 O gas at each depth.
In each plot, one soil profile sampler was installed between cotton rows. 35 ml gas samples from each soil depth were collected through the corresponding sampling port using the disposable airtight syringe. The gas sample was then injected into preevacuated 35-ml gas-tight aluminum bags (Hede Technologies, Dalian, China). In total, gas samplings were performed 13 times between 14 May and 9 November (DOY 135-314) during 2016. The N 2 O concentrations were analyzed using a gas chromatography (Agilent 7890A, Agilent Technologies, Santa Clara, CA) equipped with an electron capture detector.
The effluxes of N 2 O within soil profiles were calculated based on Fick's law using the following equation (Marshall, 1959).
; d c is the difference of N 2 O concentration in the air between two soil depths, d z is the distance between two soil depths (m). When d c is the difference between the soil at 5 cm depth and air N 2 O concentrations in the atmosphere, it is used to assume the N 2 O emission rate of the soil surface. The efflux gradient between two soil depths was used to characterize N 2 O production rates at different soil layers using the following equation (Yoh et al., 1997;Kusa et al., 2010;Nan et al., 2016).
where P i and q i are the N 2 O production rate (g m −2 s −1 ) and efflux (g m −2 s −1 ) of each soil layer, respectively. Soil gas diffusion coefficient D p was estimated using the SWLR (structure-dependent water-induced linear reduction) model (Moldrup et al., 2013).
where D 0 is the gas diffusion coefficient (m −2 s −1 ); ε is the soil airfilled porosity (m 3 m 3 ); Φ is the soil porosity (m 3 m 3 ); C m is the media complexity factor in the SWLR model, and Moldrup et al. (2013) recommended a value of 2.1 for C m in intact soils after comparing several prediction models.
where ρ s is the average particle soil density (2.65 g m −3 ); and θ is the soil bulk density (g m −3 ) and soil volumetric water content of each soil layer. The diffusion coefficient D 0 was calculated based on temperature and pressure using the following equation (Campbell, 1985): where T and P are the temperature (°C) and air pressure (Pa), respectively; D s is 1.43 × 10 −5 m 2 s −1 , that is the diffusion coefficient of N 2 O in free air at the reference temperature (273.15 K) and reference air pressure P 0 (1 atm) (Pritchard and Currie, 1982). P values for each sampling day were derived from a weather station in the field. The growing-season cumulative N 2 O emissions (ƩN 2 O, g N 2 O-N ha −1 ) were calculated by summing up the daily average emissions calculated from the concentration gradient method. Linear interpolation was used to estimate the missing values where a sampling was not conducted. Similar method was used to calculate the cumulative N 2 O production (ƩN 2 O P , g N 2 O-N ha −1 ) from different soil layers.

Soil sampling and analysis
In each plot, three soil samples of 0-20 cm were collected on the day of gas collection and mixed as one soil sample to measure the concentrations of NH 4 + -N and NO 3 − -N over the experiment. Soil temperature and volumetric water content (VWC) were measured using a sensor and data were collected using a data logger. For details about the sensor and logger see Kuang et al. (2018). The installation positions of sensors were at 5, 15, 30, and 60 cm under the drip tape in the drip irrigation treatment and the corresponding location in the flood irrigation treatment. Soil WFPS was calculated as follows: Where, in each layer, B D is bulk density (Mg m −3 ) and P D is particle density (assumed 2.65 Mg m −3 ).

Statistical analysis
A two-way analysis of variance was used to test the main and interactive effects of fertilizer treatment and irrigation method on ƩN 2 O and ƩN 2 O P (PROC MIXED). The N resource and irrigation were considered as fixed factors, while plot replicates were considered as random factors. Means of treatments were compared using the least significant differences when the main or interactive effects were significant. The relationship between N 2 O concentration and temperature, WFPS, air content at 5, 15, 30, and 60 cm depths, as well as NO 3 − -N and NH 4 + -N of 0-20 cm top soil were examined by regression analysis. Similar regression analysis was used to examine the relationship of N 2 O flux rates between the concentration-gradient method (GM) and the closed-chamber method (CM). The surface N 2 O flux rates based on the closedchamber method were previously reported by Kuang et al. (2018) and used in this study for comparison between the two methods. The normality and homogeneity of variance were checked before analysis. Differences were considered as significant at p < 0.05. All analysis were performed using the Statistical Analysis Software package (SAS Institute, 2011).

Environmental and soil conditions
Irrespective of irrigation type, soil temperature at 5 cm soil depth followed a similar trend as for air temperature, gradual increased from April to July and then decreased (Figure 1). The annual total precipitation was 50 mm in 2016. In comparison, the total water addition was 593 mm and 982 mm for drip and flood irrigation plots, respectively, which accounted for 92%-95% of the total water inputs. Soil WFPS in both drip and flood irrigation soils showed large fluctuations in response to irrigation and rainfall events, which ranged from 7.4%-43.1% and 6.4%-46.7% for drip and flood irrigation, respectively. For drip irrigation plots, soil WFPS increased with irrigation events at 5, 15, 30, and 60 cm depths, but the peaks at 5 and 15 cm depths tended to be larger than those at 30 and 60 cm. For flood irrigation plots, soil WFPS at 5, 15, 30, and 60 cm layers increased with irrigation events, with similar peaks in different depths, but the rate of water decline in deeper soils was generally slower than that in shallow soils.

N 2 O efflux rate
The soil N 2 O efflux rate at each depth varied with fertilization and irrigation treatments. The peak N 2 O efflux rate was higher under flood irrigation (42.91 μg N m −2 h −1 ) than under drip irrigation (20.63 μg N m −2 h −1 ) treatment, and all fertilization treatments had higher N 2 O efflux peaks than the control (drip: 18.73 μg N m −2 h −1 , flood: 36.13 μg N m −2 h −1 ) under both drip and flood irrigation conditions. Across all treatments, the 0-5 cm soil depth had the highest N 2 O efflux rate, ranging from −4.60-42.91 μg N m −2 h −1 (Figure 3).

N 2 O production rate
The N 2 O production rates of different soil depths were calculated from the N 2 O efflux rates of two adjacent layers. Overall, the N 2 O production rate in the profile decreased with increasing soil depth irrespective of fertilizer and irrigation treatments, with the 0-5 cm layer having the highest N 2 O production rate (Figure 4). The N 2 O production rates in the 0-5 cm soil layer were 3.3, 4.2, 6.5, and 8.0 μg N m −2 h −1 for Control, Urea, U + M, and Manure treatments under drip irrigation, respectively.
Cumulative N 2 O production was higher in the 0-5 cm soil layer than other depths under all treatments, and average cumulative N 2 O production at 5, 15, 30, and 60 cm layers were 248, 27, −10, and −3 g N ha -1 , respectively ( Table 2). Calculation of the contribution of cumulative N 2 O production in each soil depth to Frontiers in Environmental Science frontiersin.org surface N 2 O emissions based on the concentration-gradient method showed that the 0-5 cm and 5-15 cm depths contributed to all surface N 2 O emissions, with the 0-5 cm soil depth contributing 87%-100% of the surface emissions. The contribution of cumulative N 2 O production from the 0-5 cm soil layer to surface N 2 O emissions was significantly affected (p < 0.05) by irrigation treatment, being 99% and 87% for drip and flood irrigation conditions, respectively.

Comparison of surface N 2 O emission and flux between GM and CM
Using the data of surface N 2 O flux measured by the closedchamber method reported by Kuang et al. (2018), we compared the surface N 2 O emission and flux between CM and GM (Table 3). Results showed a non-significant (p = 0.128) relationship of N 2 O flux between CM and GM. Fertilizer treatments significantly affected cumulative surface N 2 O emissions based on these two methods (p < 0.05), with a numerical trend of Control < Urea < U + M < Manure (Table 3). There was no significant difference in the effect of the irrigation method on cumulative N 2 O emissions based on the concentration-gradient method, but the closed-chamber method showed cumulative N 2 O emissions significantly higher under drip irrigation than flood irrigation (p < 0.05).

Relationship between soil profile N 2 O concentration and environmental factors
Pearson correlation analysis was used to investigate the relationships of soil profile N 2 O concentration with soil temperature, WFPS, soil pore air content and inorganic N content of 0-20 cm depth. Results showed a significantly (p < 0.05) negative relationship between soil profile N 2 O concentration and soil temperature (Table 4). A significantly (p < 0.05) negative relationship was also shown between soil N 2 O concentration at 15, 30, and 60 cm soil depth and NO 3 − -N content at 0-20 cm soil depth, however, N 2 O concentration at 5 cm depth was significantly and positively correlated with NH 4 + -N content at 0-20 cm depth (p < 0.05). In addition, there was a significantly (p < 0.05) positive correlation between N 2 O concentrations in different soil layers. 4 Discussion 4.1 Low N 2 O emissions from cotton fields in the arid region are due to the generally low N 2 O production and the transmission of N 2 O from the near-surface soil to the subsoil In this study, soil N 2 O concentrations were generally low in all soil layers under both drip and flood irrigation conditions, with average N 2 O concentrations ranging from 333-359 nL L −1 in the profile, much lower than values reported in previous studies for other terrestrial ecosystems (Wang et al., 2013;Nan et al., 2016;Zhou et al., 2016;Wang et al., 2018;Yao et al., 2018). During the observation period, the cumulative N 2 O production from different soil layers ranged from −25-306 g N ha −1 , which were also lower than previous reported results by Yao et al. (2018). These results indicate low N 2 O concentrations and production in sandy soils, which were likely associated with the low soil moisture and C content (Kuang et al., 2018;Yin et al., 2019).
Soil surface N 2 O emissions are a net gas exchange between soil and atmosphere and N 2 O concentrations at different depths reflect the combined effects of N 2 O processes in the soil. Several previous studies have found that N 2 O concentrations increased with

FIGURE 1
Daily water input (irrigation + rainfall), air temperature and soil temperatures at 5 cm depth, and water-filled pore space (WFPS) at 5, 15, 30, and 60 cm depths under drip and flood irrigation.

Frontiers in Environmental Science
frontiersin.org increasing soil depth, with high N 2 O concentrations in the subsoil but a low contribution to surface N 2 O emissions (Wang et al., 2013;Wang et al., 2018;Yao et al., 2018), mainly because N 2 O produced in the subsoil may undergo complete denitrification to N 2 during upward transport (Gao et al., 2014;Kuang et al., 2019). In contrast, the N 2 O concentrations through the soil profiles in this study were all low, and the overall distribution pattern was uniform or higher in the surface soil layer than in the bottom layer. These results implies that the deep-depth soil may be a sink for these processes in cotton fields in the arid zone, and this inference can be validated by the fact that the cumulative N 2 O production in the 30 and 60 cm layers was mostly negative (Table 1). Furthermore, some previous studies have also found higher N 2 O concentrations in the near-surface soil than in the subsoil with a high soil moisture (Sotomayor and Rice, 1999). The low soil N 2 O concentrations also indicate that sandy soils were conducive to nitrification process under the aerobic condition, and denitrification process under the anerobic conditions. These biological processes need involvements of soil microbe and organic C, which are generally low in sandy soils as in the current study. Low water holding capacity and high hydraulic conductivity in sandy soils could have decreased WFPS to a level less conducive to denitrification. In addition, the high permeability of sandy soil could facilitate the vertical transfer of N 2 O between depths, and thus a less differentiated concentration gradient.

Topsoil is the main source of N 2 O emissions from this cotton field in the arid region
The topsoil (0-15 cm) contributed to all the surface N 2 O emissions in the present study, which is consistent to previous studies (Nan et al., 2016;Yao et al., 2018;Li et al., 2021). There might be two reasons for the high N 2 O production rate of topsoil. The first reason is that the surface soil contains a high concentration of organic carbon and nitrogen due to the suitable temperature and humidity, which stimulate litter and root decomposition as well as the mineralization of organic fertilizer and urea. Second, the higher temperature and humidity in the topsoil promote the occurrence of microbial processes that produce N 2 O . According to Li et al. (2021), surface N 2 O emissions were mostly associated with the topsoil of 0-15 cm under a drip irrigation cotton field. Nan et al. (2016) also reported that surface N 2 O emissions Frontiers in Environmental Science frontiersin.org originated entirely from 0-30 cm layer. These may be due to the high microbial activity of soils in the 0-30 cm layer, which was dominated by N fertilizer-induced N 2 O production (Wang et al., 2013). In contrast, the study's findings that subsoil may contribute negatively to surface N 2 O emissions are primarily attributable to the higher gas diffusion coefficient caused by the lower water retention capacity of the sandy soil in the region. This assumption is supported by the fact that N 2 O concentrations in different soil layers correlated positively with one another (Table 4). These results indicated that topsoil was the main source of surface N 2 O emissions, and the subsoil played a completely different role in N 2 O production and consumption in different studies. Thus, the role of the subsoil should be considered when studying the processes of N 2 O emissions from the more permeable sandy soils. Over the study period, the cumulative N 2 O emissions in the topsoil of 0-5 cm contributed 87%-100% of the surface N 2 O emissions. It showed a significant positive correlation (p < 0.05) with soil NH 4 + -N content but a non-significant negative correlation (p > 0.05) with soil NO 3 − -N content at 0-20 cm depth (Table 4). These results imply that the nitrification process may be the main source of N 2 O production in 0-5 cm soils, due to the low WFPS (10%-40%) of 0-5 cm soil layer and its decreased sharply after irrigation or rainfall events (Figure 1), which create a suitable condition for nitrification. Similarly, Bateman and Baggs (2005) also found that nitrification was the main pathway of N 2 O production in sandy loam soils with 35%-60% WFPS using the 15 N labeling method.

4.3
The key soil layers effected by nitrogen management and irrigation which on N 2 O production are different Both GM and CM results showed that the addition of manure significantly increased the N 2 O flux, which was consistent with the result reported by Zhou et al. (2017). Manure addition can increase soil C availability and also the activities of associated N-cycling microbes, and consequently stimulate the nitrification and denitrification processes (Lessard et al., 1996). In addition, some studies have shown that simultaneous inputs of C and NH 4 + into the soil, can promote the emission of N 2 O more than the addition of NH 4 + alone (Bergstrom et al., 1994). In the current study, through the N 2 O production and consumption in the profile, it can be found that the addition of manure significantly increased N 2 O production in the 0-5 cm layer. That is, the influence of the N source on the N 2 O production of the profile is mainly manifested in the surface layer (0-5 cm). This may be related to the mixing of manure with surface soil (0-20 cm) in this study. The topsoil has higher microbial activity than the subsoil (Uchida et al., 2011). Abundant organic C compounds and available nitrogen in manure provide sufficient Frontiers in Environmental Science frontiersin.org energy and substrates for the nitrifying and denitrifying reactions of surface soil microbial biomass which was promoted N 2 O emission (Lentz et al., 2014). Furthermore, the correlation between N 2 O (0-5 cm) production and NH 4 + -N, shows that the increase in N 2 O flux caused by manure addition is the result of nitrification.
Using both individual experiment and meta-analysis, previously reported that flood and furrow irrigation resulted in significantly higher N 2 O emission than drip irrigation (Sánchez-Martín et al., 2008;Kuang et al., 2021). In this study, GM was used to find that flood irrigation significantly increased the cumulative production of N 2 O in the 5-15 cm compared with drip irrigation, further confirming that the effect in the irrigation method on N 2 O production was mainly expressed in the sub-surface layer. This may be associated with the disparities of soil profile water status  For each treatment factor, means within a column followed by the same letter are not significantly different at p < 0.05.
Frontiers in Environmental Science frontiersin.org caused by different irrigation methods. Compared with drip irrigation, the water content in the whole profile under flood irrigation was higher and decreased more slowly, which provided a optimal moisture condition for the production of N 2 O in the profile . In addition, the higher risk of N leaching under flood irrigation may have increased the nitrogen content at 5-15 cm depth, enhancing nitrification/denitrification processes and thus increasing N 2 O production. In this study, no significant interaction between irrigation and N source treatments on N 2 O production in each soil layer was found. Although a discrepancy exists in the N 2 O emission estimated by the CM and GM, the trend of cumulative N 2 O emissions under different treatments is consistent. Therefore, GM's ability to provide abundant information on profile N 2 O production, reduction, and diffusion is commendable.

Conclusion
In the current study, soil N 2 O concentrations were generally low in all soil layers under both drip and flood irrigation conditions, with average N 2 O concentrations ranging from 333-359 nL L −1 through the soil profile. Such low N 2 O concentration and efflux in sandy soils were mainly attributed to the generally low N 2 O production due to low soil moisture and C availability. Soil N 2 O concentrations decreased with increasing soil depth, and had similar seasonal patterns in different soil layers, indicating that the subsoil was a N 2 O sink, which also explained the low N 2 O emissions in the current study. Additionally, the topsoil (0-15 cm) contributed all the surface N 2 O emissions, with a contribution of 87%-100% at a topsoil of 0-5 cm. The increase of soil moisture and C, N availability under N source and irrigation treatments were the main factors influencing the N 2 O production in the 5 and 15 cm soil layers. N inputs through synthetic fertilizer or manure significantly increased N 2 O concentration and production at 0-15 cm. Flood irrigation resulted in higher N 2 O concentration and production compared to drip irrigation. The results confirm that soil water and nitrogen management are important drivers of N 2 O production and diffusion in soil profiles of croplands in arid region. Use of nitrification inhibitor to slow N transformation process in soils can be an effective strategy to reduce N 2 O production and emissions in arid regions. Future research is also needed to incorporate fertilizer into drip irrigation strategy, i.e., fertigation, to develop effective strategies for enhancing crop productivity while mitigating GHG emissions. For each treatment factor, means within a column followed by the same letter are not significantly different at p < 0.05.