This study was carried out under the supervision of the department of Soil and Water Conservation at Dingxi Gansu Province (35°34′53′′ N, 104°38′30′′ E longitude with an altitude of 2000 m above sea level), Northwestern region of China (Figure 1). The research area is a typical hilly and gully area of the Loess Plateau. The climatic condition of study zone is semi-arid, with a large annual temperature difference. The average annual temperature is 6.9°C, the average annual precipitation is 400 mm, mainly concentrated in July–September, the average annual solar radiation is 5,930 MJ m⁻² with 2,480 h of sunshine per year, a frost-free period of 140 days. The soil of study areas is typical loess with sandy loam in texture. The area is dry with water deficient and the vegetation is scant. The natural vegetation is dominated by grasses such as Gramineae, Leguminosae, Compositae, mostly Medicago sativa, Onobrychis viciifolia and a small number of shrubs are distributed such as Caragana Korshinski, Hippophae rhamnoides. The arbor species are mainly Platycladus orientalis, Picea asperata, Xanthoceras sorbifolium, etc. The main crops of this area are Zea mays, Solanum tuberosum, Linum usitatissimum, Triticum aestivum.

Based on the field investigation of the ecological environment characteristics and vegetation characteristics of the study area we took six different land use patterns viz., PA (Picea asperata), HR (Hippophae rhamnoides), MS (Medicago sativa), NT (no-tillage in wheat field), T (Conventional tillage in wheat field), on Loess Plateau in Central Gansu (Table 1; Figure 2). Three small plots were randomly selected in each experimental area, and the size depended on the vegetation type: wheat field, grassland, Medicago sativa grassland 4 m × 6 m, Hippophae rhamnoides woodland 10 m × 10 m, Picea asperata woodland 20 m × 20 m. The selected wheat field was the local spring wheat “Ganchun 35” as the test variety, which was sown on March 25, 2021 and harvested on August 5. The seeding rate of each plot was 187.5 kg/hm², and the row spacing was 25 cm. A combination of 150 kg/hm² of diammonium superphosphate (N + P₂O₅) and 62.5 kg/hm² of urea was used as the basal dose. At the same time, a fixed gas extraction area of 0.5 m × 0.5 m was randomly set up in each fixed sample plot to collect and determine N₂O flux under different land use patterns.

The disturbed soil samples from different experimental treatments (GL, PA, HR, MS, T and NT) were collected with an auger of 5 cm diameter, from March 2021 to February 2022 for the determination of soil properties. Five soil samples were taken from each individual plot and placed in a plastic bag and transported to laboratory for physical chemical analysis. The soil temperature was monitored by using the EM50 data collector after every 30 min. Soil water content, soil organic carbon, total nitrogen, nitrate nitrogen, ammonium nitrogen, microbial biomass nitrogen and soil pH electrical conductivity were determined in laboratory after passing soil samples through a 2 mm sieve. The oven-drying method was used for gravimetric soil water content determination (Défossez et al., 2021). The soil pH was measured by Potentiometric method (soil water ratio = 1:2.5) (Mao et al., 2016). The core sampler method was used for determination of soil bulk density (Lu et al., 2019). Soil organic carbon (SOC) was determined by using the Walkley-Black dichromate oxidation method (Nelson, 1982). Soil TN was determined by the oxidation-external heating titration method (Huang et al., 2021), and soil microbial nitrogen (MBN) was determined using a chloroform fumigation-H₂SO₄ leaching method (Chen et al., 2021). The soil NO₃⁻-N and NH₄⁺-N were determined by the MgO-Devarda’s alloy distillation method (Yang et al., 2014).

N₂O emissions from soil shall be measured every half month from March 2021 to February 2022. Two weeks prior to starting measurements, three stainless steel bases (50 cm × 50 cm × 20 cm) were permanently installed in the soil at each site with at least a 5 m buffer zone between each steel base. The upper part of each base had a collar (3 cm) to support the sampling chamber. Located at the top of each open-bottom opaque stainless steel chamber (50 cm × 50 cm × 50 cm) were two fans powered by 12 v batteries used to mix the air inside the chamber during the N₂O flux measurements. A rubber seal strip was attached to the bottom perimeter of each chamber to provide an airtight seal when placed over the collar on the base in the soil. As done in previous studies (Ma et al., 2018), all samples were collected between 09:00 a.m. and 11:00 a.m. (local time) on the same day each week when the soil temperature was close to the daily mean soil temperature (Zou et al., 2005; Yao et al., 2013). After the chamber was closed, air samples (five in total) were taken from inside the chamber at 8 min intervals (at minutes 0, 8, 16, 24, and 32) using a 100 ml polypropylene syringe equipped with a three-way plug valve. The air samples were transferred to pre-evacuated E-Switch aluminum foil composite film gas sampling bags (Shanghai Shenyuan Scientific Instruments Co., Ltd., Shanghai, China) via the three-way plug valve. Within 48 h, the samples were returned to the laboratory for analysis of N₂O concentrations using a gas chromatograph (Agilent 4890D, Agilent Technologies, Wilmington, Delaware, USA). The ratios of each set of 5 samples were linearly fitted to the corresponding sampling intervals. When the linear regression coefficient R² > 0.75, it is regarded as valid data and used to calculate the emission flux of the target gas.

Calculation of the emission flux of N₂O:

In the formula: F is the emission flux of N₂O, μg·(m²· h) ⁻¹, ρ is the density of N₂O in the quasi-state (kg·m⁻³), V is the adequate space volume in the airtight box (m³), A is the water surface area covered by the closed box (m²), ΔC is the concentration difference of N₂O gas (expressed as the volume fraction of air), Δt is the sampling interval (h), T is the temperature of the closed box during sampling (°C).

Calculation of N₂O cumulative emissions:

In the formula: M is the cumulative N₂O emission (kg· hm⁻²), F is the emission flux of N₂O, μg·(m²· h) ⁻¹, t_i + 1−t_i is the time interval of the ith and i + 1 sampling (d), n is the total number of measurements during the observation period.

The study data obtained from the research field were tested at 5% probability level with one-way factor interaction ANOVA by using an appropriate SPSS 25 (IBM Corp., Chicago, IL, USA) computer software program. The significant differences among different experimental treatments and their interaction were compared with an LSD test. The relationships amongst different soil properties were analyzed using Pearson’s correlation coefficient. The study data is presented as the mean values of three replications with standard deviation. Furthermore, principle component analysis (PCA) was done in order to assess the multivariate variability introduced by the different treatments for soil properties and N₂O emission in the soil system.

Seasonal soil temperature variation under six different land use patterns was consistent. From the beginning of the growing season, the soil temperatures fluctuated upward as arial temperatures gradually warmed (Figure 3). The soil temperature peaked from mid-July to mid-August during summer whilst it declined in December during winter. Soil temperature under GL treatment was highest and it was lowest under PA treatment. The mean soil temperature under the GL, PA, HR, MS, T and NT treatments was 11.66, 8.78, 9.31, 10.16, 9.68 and 10.01°C, respectively.

Land use patterns significantly affected soil properties (Table 2). Compared with GL treatment, Picea asperata woodland significantly increased soil bulk density, soil water content and soil SOC and decreased soil pH value. Soil pH values showed that the soils of experimental site were alkaline, and the average pH of wheat fields was similar and significantly higher than that of the control sites, with the highest in T treatment and the lowest in PA treatment. Soil water content was the highest in PA treatment and lowest in T treatment. The soil SOC contents under the six land use patterns were PA > GL > T > NT > HR > MS, being significantly higher under Picea asperata woodland than other land use patterns. There was no significant difference between CT and NT treatments in the wheat field.

Soil nitrogen is an important indicator reflecting soil fertility and biological activity. It can be seen from Figure 4 that among the six different land use patterns, the soil TN content of the GL plot was significantly lower than that of the other five vegetation types (p < 0.05), and the specific trend was PA > HR > MS > NT > T > GL. Compared with GL, the contents of TN increased by 41.44, 30.63, 25.45, 15.99 and 11.04% under PA, HR, MS, NT, and T treatments, respectively. There was no significant difference was noted between HR, MS, T and NT treatments; however, significant difference (p < 0.05) was seemed between PA treatment and other treatments. Soil MBN content was significantly affected between different land use patterns (Figure 4B). The PA treatment had significant higher MBN (28.28 mg kg⁻¹) whilst T treatment had lowest (18.03 mg kg⁻¹) MBN value. The PA, HR and MS increased MBN by 38.36, 22.11 and 9.05%, respectively. Our results showed that Picea asperata woodland increased TN and MBN accumulation.

The content of soil NO₃⁻-N (32.02 mg/kg) was the highest under T treatment and the lowest in treatment GL (26.85 mg/kg), followed the trend of T > NT > MS > HR > PA > GL. In addition, the content of NH₄⁺-N under GL treatment was lower than that of other land use patterns. Compared with GL treatment, the NT, T, MS, HR, and PA treatments increased soil NH₄⁺-N by 23.8, 16.61, 14.04, 11.9 and 8.65%, respectively. In general; compared with GL, the contents of NO₃⁻-N and NH₄⁺-N in soil under other land use patterns increased with varying degrees especially in conventional and no-tillage system in wheat field.

Our study data depicted apparent seasonal variations in N₂O emission. The highest N₂O emission risen during summer months being peak occurring in July under CT followed by NT, HR, GL, MS and PA. In addition, there was no significant difference in soil N₂O emissions under the different land use patterns during autumn and winter months (Figure 5).

All land use patterns served as nitrous oxide emitters during the study period. Our results showed that the T treatment had highest cumulative N₂O flux value whilst PA treatment had lowest cumulative N₂O flux value. The order of cumulative N₂O flux under all treatments was T > NT > HR > GL > MS > PA (Figure 6). Compared with PA, the T, NT and HR treatments increased cumulative N₂O emission by 113.64, 91.02 and 16.66%, respectively.

According to Jolliffe cut-off value the Principle component analysis (PCA) permitted for isolating five principal components. The observation point made by a collaboration of PC1 and PC2 shows the general variance defined by the five main components in accordance with PCA analysis.

In the PCA analysis of nine variables (viz., ST, SWC, pH SOC, MBN, TN, N₂O, NO₃⁻-N and NH₄ ⁺-N), PC1 and PC2 were extracted with eigenvalues (>1) and explained 63.8% of the total variance. However, PC3, PC4, and PC5 do not permit the supplementary information addition; that is why they are not involved. The maximum loadings of PC1 comprise 42.3% of the total variance and in PC2 the higher loadings of 21.5% of the total variance were detected (Figure 7).

Our study showed significant positive correlation between N₂O emission and soil and environmental factors. Soil temperature had significant positive correlation with N₂O emission. Furthermore, soil water content, soil pH, and soil NO₃⁻-N, NH₄⁺-N, TN and MBN were significantly positively correlated with N₂O emission. Additionally, a negative correlation was observed between N₂O emission with soil organic carbon. Fluctuations in N₂O emission appeared to be described by dissimilarities in soil temperature, moisture content and nitrogen contents with data variance 95% described by these soil and environmental variables (Figure 8).

Land use is one of the essential components of an ecosystem and its changes have a significant impact on soil properties (Deng et al., 2018). Previous study by Yan et al. (2021) have shown that there are overall soil pH differences among land use patterns and forest land has the lowest pH. Another study by Wang et al. (2020) found that the forest soil system has better water holding capacity and water supply capacity, and the SOC content in the forest is significantly higher than shrubs and grasslands. Our results found that compared with Grassland, Picea asperata woodland significantly increased soil bulk density, soil water content and soil SOC content, and decreased soil pH because arbor forest land was richer in vegetation, litter and root systems There are more secretions, and the decomposition process of litter will cause soil acidification (Qiao et al., 2014). At the same time, forest vegetation can also change the forest environment, reduce solar radiation and temperature differences and increase soil moisture (Özkan and Gökbulak, 2017), Creating a stable environment for litter decomposition. In addition, due to the large degree of human disturbance in wheat fields and shrub grasslands, the vegetation community structure is relatively simple, and the litter and root densities are lower than those of arbor fields. Therefore, Picea asperata woodland can increase the soil bulk density, organic matter and water content and reduce the soil pH value.

Different land use patterns can affect the content and distribution of nitrogen components in the soil to varying degrees (Cheng et al., 2013). Studies have shown that nitrogen-fixing tree species in forests can increase soil nitrogen stocks and promote soil nitrogen cycling (Rothe et al., 2002). This study showed that the soil N content of the five land-use patterns increased to varying degrees compared with that of the Grassland, with the total N and microbial N content of the Picea asperata woodland being significantly higher than that of the shrub grassland and wheatland. On the one hand, the Picea asperata woodlands have lush vegetation growth and are accompanied by a large number of herbs, so the Picea asperata woodlands provide a large amount of litter, animals, and microbial residues for the soil surface, which can effectively speed up the decomposition rate of vegetation litter and promote the return rate of nutrients. (Anurag et al., 2019). On the other hand, because the biomass of woodland vegetation is mainly concentrated in the above-ground part, and the surface soil is well-aerated and nutrient-rich, it is more conducive to soil microbial activity. In addition, the ripe harvest of crops will lead to the loss of a large amount of organic matter in the soil, and various tillage measures led to soil system changes, which accelerate the mineralization and decomposition of soil organic matter (Shi et al., 2013), thus leading to TN and MBN accumulation in forest soils. The content is significantly higher than that of agricultural grassland. The content of ammonium and nitrate nitrogen in wheat field was significantly higher than in the other treatments. The specific performance was that the soil ammonium nitrogen content was the largest under the NT treatment and the nitrate nitrogen content was higher under the T treatment. This is because the human disturbance in wheat fields is higher than that in forest and grassland, and the application of basal fertilizer during sowing will lead to the accumulation of ammonium and nitrate nitrogen in the soil. In addition, NT treatment reduced soil compaction and soil disturbance benefits the nutrient accumulation and crop root growth whereas plough layer under CT increased the soil compaction (Fiorini et al., 2020). The high NO₃⁻- N content of the soil surface under conventional tillage is due to the destruction of the soil surface layer by conventional tillage, which results in the loss of soil nutrients and reduced moisture content, making it easier to convert NH₄⁺-N to NO₃⁻- N which in turn leads to an increase in soil NO₃⁻- N content (Pisani et al., 2017).

The N₂O flux varies significantly under different land use patterns. In this study, the average N₂O flux from agricultural soils was more significant than that from grassland, which is consistent with the findings of Lin et al. (2009). The cumulative N₂O flux of farmland soil was highest, possibly due to the increase of nitrogen content by nitrogen fertilizer application in farmland, thus promoting N₂O emission. Furthermore, the systematic plowing of farmland can promote mineralization to generate ammonium and nitrate nitrogen thus increasing soil nitrogen accumulation. When soil temperature and humidity conditions are suitable, a large amount of N₂O is released. However, there is also a study by Schaufler et al. (2010) who declared that the grassland had more N₂O emission than farmland. The high activity of microorganisms in grassland stimulates the accumulation of soil nitrogen and the dense roots of grassland and litter on the ground are the reasons for the high N₂O emission from grassland. The N₂O flux is higher in spring and summer but lower in autumn and winter. The first emission peak of N₂O occurred around April 15, specifically for the T treatment. This is due to the application of basal fertilizer at sowing, which starts to take effect and allows some of the nitrogen to enter the soil and the soil to emit some of the N₂O through respiration but due to the low temperature and weak microbial activity of the soil at this time, the N₂O emissions are relatively low and therefore a small emission occurs. From July to August, the soil N₂O emission was more under all treatments being highest under T treatment because during summer the hydrothermal conditions are more good and suitable. The temperature regulates the N₂O emission from the soil through its influence on the activities of nitrifying bacteria, denitrifying bacteria and other microorganisms. In addition, the rainfall during summer months was relatively more concentrated and rainfall water infiltrated into the soil which led to an anaerobic conditions and promotes nitrification and denitrification processes. This is consistent with the findings of Pendall et al. (2010). In addition, at the end of spring wheat growth, the demand for nitrogen in the soil becomes less. At the same time, the surface litter returns nitrogen source to the soil through decomposers’ decomposition, hence the contentration of nitrogen in the soil increases, thereby enhancing denitrification (Cao et al., 2019). The N₂O emission during winter months was very low and there was almost no difference in the soil N₂O emission under different land use patterns. This might be due to the low winter temperature, the soil’s low moisture content and the inorganic nitrogen. The results of this study showed that compared with the GL treatment, the T treatment significantly increased soil N₂O emissions specifically T > NT > HR > GL > MS > PA. This is because of good and better water permeability and air permeability under conventional tillage. The conventional tillage system favors to the transformation of soil anaerobic environment to an aerobic environment, thus promotes the nitrification process to produce N₂O and also strengthen the gas diffusion and nutrient cycling, so that soil SOC, NH₄⁺-N, and NO₃⁻-N move up from deep soil to surface soil again, so that surface dead vegetation enters into the lower soil layer and accelerates its decomposition, thereby promoting the emission of N₂O in the soil (Du et al., 2022). In addition, no-tillage increases the soil capacity and reduces the gas diffusion rate, which in turn allows the N₂O produced by nitrification and denitrification to be further reduced to N₂ before it can emit from the surface soil, thus reducing N₂O emissions (Du et al., 2022). In general, the woodland had lowest N₂O emissions, probably due to unfertilized woodland and also might be taller plants, denser plants growth, shady conditions throughout the year and relatively low soil temperature which reduces evaporation and leaching losses.

N₂O emission is mediated by microbial nitrification (autotrophic/heterotrophic nitrification) and denitrification (Bond-Lamberty et al., 2004; Guetlein et al., 2018). Nitrifying (ammonia-oxidizing archaea, AOA; ammonia-oxidizing bacteria, AOB) and denitrifying genes [nitrite reductase (encoded by nirS/nirK)], are considered to be the rate-limiting step in nitrification and denitrification, respectively, which are main process in the production of N₂O (Jones et al., 2014; Kits et al., 2019). At the same time, The environmental factors such as soil temperature and soil moisture influence on soil N₂O emissions. Our correlation analysis showed a very significant positive correlation between soil N₂O emissions and soil temperature. This is because of the nitrification and denitrification processes that can be carried out from 5°C to 35°C. The soil temperature regulates N₂O emissions mainly by influencing the structure of the nirS-type denitrifying microbial community, and the activity of denitrifying bacteria increases by 1.5 to 3.0 times for every 10°C increase in temperature (Xing et al., 2021). On the other hand, the soil temperature also increases the respiration of microorganisms in the soil which produces the anaerobic conditions for denitrifying organism’s resultant in higher N₂O emission (Braker et al., 2010). This study also showed that soil N₂O emissions were significantly and positively correlated with soil water content, pH, and NO₃⁻-N, TN and NH₄⁺-N which is consistent with the results of previous study (Qin et al., 2021). The higher soil moisture mainly affects the community structure of nirK and nosZ gene-containing microorganisms, stimulating the growth of some microorganisms in small abundance, including Bradyrhizobium with the nirK gene and Tardiphaga with the nosZ gene, thus regulating the emission of N₂O (Qin et al., 2021). In addition, an increase in soil moisture content also increases the enrichment of effective soil carbon and oxygen, affecting the distribution of oxygen concentrations in the soil and the activity of nitrifying and denitrifying microorganisms, which in turn causes N₂O emissions (Hu et al., 2015). Moreover, the soil pH can directly regulate N₂O emissions by affecting biological and abiotic processes (Zaman et al., 2012). It is generally believed that the pH in between 6 and 8 is most favorable for denitrifying bacterial activity. Soil microbial activity significantly enhanced with the increasing pH value, leading to enhanced soil denitrification and increased N₂O emission (Zaman et al., 2012). Moreover, increasing soil TN, nitrate nitrogen and ammonium nitrogen content will provide a good nitrogen source for nitrification and denitrification. It has been found that ammonia-oxidizing bacteria start to grow immediately after the addition of a nitrogen source. Ammonia-oxidizing bacteria (AOB) and ammonia-oxidizing archaea (AOA) can oxidize ammonia and produce N₂O (Kou et al., 2015; Dong et al., 2018). NO₂⁻ from soil denitrification, which can accumulate under certain conditions also has a large potential for N₂O emissions (Giguere et al., 2017). At the same time, available nitrogen in soil provides an electron donor for denitrifying microorganisms, promoting the occurrence of denitrification, thus increasing N₂O emissions (Cao et al., 2019).