E ﬀ ect of Nitrogen Fertilizer on Soil CO 2 Emission Depends on Crop Rotation Strategy

: Developing environmentally friendly and sustainable nitrogen (N) fertilizer management strategies is crucial in mitigating carbon dioxide (CO 2 ) emission from soil. How N fertilizer management practices inﬂuence soil CO 2 emission rates under di ﬀ erent crop rotations remains unclear. The aim of this study was to assess the impact on soil CO 2 emission and soil physicochemical properties of three N fertilizer treatments including traditional rate (TF), optimized rate (0.8TF), and no fertilizer (NF) under three di ﬀ erent crop rotation treatments: wheat-fallow (WF), wheat-soybean (WS), and wheat-maize (WM) over two years in a ﬁeld experiment in northwest China. The rates were 5.51, 5.60, and 5.97 µ mol · m − 2 · s − 1 of mean soil CO 2 emission under the TF, 0.8TF, and NF treatments, respectively. Mean soil CO 2 emission rates were 21.33 and 26.99% higher under the WM rotation compared with the WF and WS rotations, respectively. The WS rotation showed higher soil nutrient content and lower soil CO 2 emissions, and reduced fertilizer application. Importantly, soil organic carbon (SOC) concentration in the topsoil can be maximized by including either a summer legume or a summer maize crop in winter wheat rotations, and by applying N fertilizer at the optimal rate. This may be particularly beneﬁcial in the dryland cropping systems of northern China.


Introduction
Agricultural practices contribute significantly contribute to global climate change, and agricultural soil is a major source of greenhouse gases globally [1]. Agriculture is also considered to increase soil organic carbon (SOC) and net primary production of plants during the growth period to mitigate net greenhouse gas emissions from soil [2]. CO 2 emissions from the soil are determined by various environmental factors. (1) Physical properties, such as the activity, communities and function of soil microorganisms [3,4], the quality and amount of organic matter in the soil [5], and texture [6]; (2) environmental conditions of soil such as moisture, pH value, ventilation and temperature [6,7]; and (3) management practices such as land-cover types [8,9], tillage [10], crop rotation [11], irrigation and fertilization [12,13], have been identified as the most important factors influencing soil CO 2 emissions. Because soil is a living biological system, CO 2 emission is a complex problem, and limited information exists for different rotation and nitrogen (N) fertilizer treatments. mean air temperature of 12.9 °C and an average annual rainfall of 630 mm. The warmest time occurred from July to September, which also had the highest rainfall ( Figure 1). The saturated soil water percentage and field capacity were 42.8 and 23%, respectively. The soil texture is silt clay loam, and the soil is classified as Lou soil (anthrosol), with a bulk density of 1.49 g.cm −3 . In 2008, at the beginning of the experiment, the soil was found to contain 0.86% organic matter, 12.74 mg.kg −1 alkalihydrolysable N, 21.72 mg.kg −1 available phosphorus, and 54.52 mg.kg −1 available potassium.

Experimental Design and Management
The field experiment (established for more than 10 years) was a split-plot design (three replications) with rotation in the main plot and fertilizer treatment in the subplot ( Figure 2). The measurements were carried out from June of 2011 to June of 2013. Three rotation strategies that are common in the uplands of this region were used in this study. These included a winter wheat and summer fallow (WF) rotation, a winter wheat (Triticumaestivum L.) and summer soybean (Glycine max L.Merr.) rotation (WS), and a winter wheat and summer maize (Zea mays L.) rotation (WM) ( Figure  2). The growing season of summer crops in this area is generally from June to September. The winter wheat growth period is from October to June of the following year ( Figure 1). Three fertilizer application treatments were used: (1) the local farmers' traditional practice rate (TF), in which a base fertilizer was applied once during the winter wheat season providing phosphorus pentoxide (P2O5) at 300 kg.ha −1 and urea (CON2H4) at 300 kg.ha −1 , and the WM rotation system had a top-dressing fertilizer providing 300 kg of urea (CON2H4) ha −1 , similar to the local farmers' practice rate and times; (2) optimized fertilization (0.8TF), corresponding to 80% of TF fertilizer amount; and (3) no fertilizer application (NF) (Figure 2).

Experimental Design and Management
The field experiment (established for more than 10 years) was a split-plot design (three replications) with rotation in the main plot and fertilizer treatment in the subplot ( Figure 2). The measurements were carried out from June of 2011 to June of 2013. Three rotation strategies that are common in the uplands of this region were used in this study. These included a winter wheat and summer fallow (WF) rotation, a winter wheat (Triticumaestivum L.) and summer soybean (Glycine max L.Merr.) rotation (WS), and a winter wheat and summer maize (Zea mays L.) rotation (WM) (Figure 2). The growing season of summer crops in this area is generally from June to September. The winter wheat growth period is from October to June of the following year ( Figure 1). Three fertilizer application treatments were used: (1) the local farmers' traditional practice rate (TF), in which a base fertilizer was applied once during the winter wheat season providing phosphorus pentoxide (P 2 O 5 ) at 300 kg·ha −1 and urea (CON 2 H 4 ) at 300 kg·ha −1 , and the WM rotation system had a top-dressing fertilizer providing 300 kg of urea (CON 2 H 4 ) ha −1 , similar to the local farmers' practice rate and times; (2) optimized fertilization (0.8TF), corresponding to 80% of TF fertilizer amount; and (3) no fertilizer application (NF) (Figure 2).

Soil CO2 Emission Measurements
Soil CO2 emission was measured using the open-flow dynamic method described by Kong and Wang [28][29][30]. Soil CO2 emission was measured during the growing seasons from 9 August 2011 to 24 May 2013. All measurements occurred between 9:00 and 11:00 a.m. once every 2 weeks. The

Soil CO 2 Emission Measurements
Soil CO 2 emission was measured using the open-flow dynamic method described by Kong and Wang [28][29][30]. Soil CO 2 emission was measured during the growing seasons from 9 August 2011 to 24 May 2013. All measurements occurred between 9:00 and 11:00 a.m. once every 2 weeks. The measurements were delayed by 1-2 days if it was raining on measurement day. The system consisted of a chamber and an infrared CO 2 analyzer (Huayun Co. Ltd., Beijing, China; model GXH-3010E1). The chamber was made of PVC, which was 16 cm in diameter and 15 cm in height, and was inserted 4 cm into the soil. Soil CO 2 emission was determined as described by Kong and Wang [29,30].
Soil moisture and soil temperature were measured adjacent to the chamber. Soil temperature was measured at depths of 5, 10, 15, 20, and 25 cm in the middle of the crop rows using a buried geothermometer. After crops were harvested, soil cores were taken from 0-100 cm of the topsoil at three points in each plot. Soil samples from three repetitions at the same level were mixed to form a composite sample. Soil samples were air-dried in a room, and then passed through a 0.25 mm mesh, similar to the method described by Kong and Wang [28,30]. The gravimetric soil water content was determined as described by Wang in which soil samples were dried at 105 • C for 48 h [30]. SOC constants were measured using the K 2 Cr 2 O 7 oxidation method as described by Liu [1]. Alkali-hydrolysable N was analyzed by alkaline hydrolysis [31]. Available phosphorus in the soil at the experimental site was analyzed using the Mo-Sb colorimetric method [32]. Available potassium was analyzed by flame photometry after NH 4 OAc extraction according to the method by Cambardella et al. [31].

Statistical Analyses
All data collected from the different treatments were subjected to analysis of variance (ANOVA) using SPSS software (version19.0, Amos Development Corporation, Wexford, PA, USA). Rotation and N fertilization treatments were fixed effects, and the average date of each crop growth period was a repeated measure variable in the mixed-model analysis which was used to evaluate the contribution of individual factors of fertilizer treatments and crop rotation modes. The statistical significance of variance and normality tests was determined using the data (p < 0.05). Correlations between soil moisture, soil temperature, SOC, and soil CO 2 were determined using CANOCO 4.5 (Biometris, Wageningen, Netherlands) [33].

Soil Temperature and Moisture Content
Soil temperature in the 0-25 cm soil layer was similar between N fertilizer and rotation treatments (Figure 3a,b, Figure S1). The mean soil temperature under WF, WS, and WM was 17.24, 17.21, and 17.19 • C, respectively, and the soil temperature under NF, 0.8TF, and TF, was 17.49, 17.06, and 17.09 • C, respectively. The average soil temperature under the NF treatment was higher than that under the other two fertilizer treatments during the growth period of winter wheat. The average soil temperature under the WF rotation was higher than that under the other two rotations during the summer crop growth period. During the 2011 summer crop, 2011-2012 winter wheat, and 2012 summer crop growing seasons, soil temperature significantly differed among the three rotation modes, but there were no significant differences between the three fertilizer treatments.
under the other two fertilizer treatments during the growth period of winter wheat. The average soil temperature under the WF rotation was higher than that under the other two rotations during the summer crop growth period. During the 2011 summer crop, 2011-2012 winter wheat, and 2012 summer crop growing seasons, soil temperature significantly differed among the three rotation modes, but there were no significant differences between the three fertilizer treatments. The mean soil moisture content of the WS and WM rotation modes was lower than that of the WF rotation under the same rainfall conditions (Figure 3c,d, Figure S2). Average soil moisture was 15.90, 15.43, and 15.62% under WF, WS, and WM, respectively. The average soil moisture levels were 15.72, 15.72, and 15.52% under the NF, 0.8TF, and TF treatments, respectively. Soil moisture content was 17%-21% during the summer crop growth periods over the two years, since this was the season when precipitation increased (Figures 1 and 3). Soil water content during the maturity stage of winter wheat, when wheat growth requires extra water, was 8%-15%. Mean soil moisture varied significantly under the 2011 summer crop and the 2012-2013 winter wheat among the three rotation treatments. However, N fertilizer showed no significant differences in mean soil moisture content. The mean soil moisture content of the WS and WM rotation modes was lower than that of the WF rotation under the same rainfall conditions (Figure 3c,d, Figure S2). Average soil moisture was 15.90, 15.43, and 15.62% under WF, WS, and WM, respectively. The average soil moisture levels were 15.72, 15.72, and 15.52% under the NF, 0.8TF, and TF treatments, respectively. Soil moisture content was 17%-21% during the summer crop growth periods over the two years, since this was the season when precipitation increased (Figures 1 and 3). Soil water content during the maturity stage of winter wheat, when wheat growth requires extra water, was 8%-15%. Mean soil moisture varied significantly under the 2011 summer crop and the 2012-2013 winter wheat among the three rotation treatments. However, N fertilizer showed no significant differences in mean soil moisture content.

Soil Organic Carbon (SOC)
The mean SOC content in the 0-100 cm soil layer varied under different rotations and N fertilizer treatments. Under the WF rotation, the SOC content was 5.45, 5.31, and 5.41 g/kg, under the WS rotation it was 5.53, 5.68, and 5.42 g/kg, and under the WM rotation it was 5.72, 5.47, and 5.66 g/kg for the N fertilizer treatments NF, 0.8TF, and TF, respectively ( Figure S3). After 10 years under the crop rotations and N treatments, treatments showed significant differences with respect to SOC content. The mean SOC content increased by 24.56, 20.44, and 5.58% under the MF, MS, and MW rotations mode Sustainability 2020, 12, 5271 6 of 14 over the study period in the 0-10 cm soil layer, and by 10.30, 20.12, and 20.17% under the TF, 0.8TF, and NF treatments, respectively ( Figure 4). During the 2011 summer crop period, significant differences in SOC were found between the three rotation modes in the 0-10 cm and 11-40 cm soil layers.
The mean SOC content in the 0-100 cm soil layer varied under different rotations and N fertilizer treatments. Under the WF rotation, the SOC content was 5.45, 5.31, and 5.41 g/kg, under the WS rotation it was 5.53, 5.68, and 5.42 g/kg, and under the WM rotation it was 5.72, 5.47, and 5.66 g/kg for the N fertilizer treatments NF, 0.8TF, and TF, respectively ( Figure S3). After 10 years under the crop rotations and N treatments, treatments showed significant differences with respect to SOC content. The mean SOC content increased by 24.56, 20.44, and 5.58% under the MF, MS, and MW rotations mode over the study period in the 0-10 cm soil layer, and by 10.30, 20.12, and 20.17% under the TF, 0.8TF, and NF treatments, respectively ( Figure 4). During the 2011 summer crop period, significant differences in SOC were found between the three rotation modes in the 0-10cm and 11-40cm soil layers.

Soil CO2 Emission
Soil CO2 emission under the three N fertilizer treatments and three rotation treatments followed similar trends over time ( Figure 5, Figure S4). High soil CO2 emission rates occurred when soil temperature was high, and low soil CO2 emission rates occurred when soil temperature was low ( Figure 5). The average soil CO2 emission was 5.51, 5.60, and 5.97 μmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively. The average soil CO2 emission rate under the WF, WS, and WM

Soil CO 2 Emission
Soil CO 2 emission under the three N fertilizer treatments and three rotation treatments followed similar trends over time ( Figure 5, Figure S4). High soil CO 2 emission rates occurred when soil temperature was high, and low soil CO 2 emission rates occurred when soil temperature was low ( Figure 5). The average soil CO 2 emission was 5.51, 5.60, and 5.97 µmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively. The average soil CO 2 emission rate under the WF, WS, and WM rotation modes was 5.39, 5.15, and 6.54 µmol·m −2 ·s −1 , respectively ( Figure 5). Except for during winter wheat growth in the 2011-2012 season, there were no significant differences in average soil CO 2 emission rate between the three rotation modes or between the N treatments. During the study period, mean soil CO 2 emission under the WF and WS rotations decreased with increasing amounts of N fertilizer, while under the WM rotation it increased with increasing amounts of N fertilizer. rotation modes was 5.39, 5.15, and 6.54 μmol·m −2 ·s −1 , respectively ( Figure 5). Except for during winter wheat growth in the 2011-2012 season, there were no significant differences in average soil CO2 emission rate between the three rotation modes or between the N treatments. During the study period, mean soil CO2 emission under the WF and WS rotations decreased with increasing amounts of N fertilizer, while under the WM rotation it increased with increasing amounts of N fertilizer. During the summer crop growing seasons ( Figure 5), the soil CO2 emission differed significantly (p = 0.0095 and p = 0.03 in 2011 and 2012, respectively) between the three rotation modes. The average soil CO2 emission rate of the WF rotation mode was the lowest of the three rotation modes, and the WM rotation was the highest. During the summer growth period of 2011, the average CO2 emission rates were 2.75, 4.96, and 6.12 μmol·m −2 ·s −1 under the WF, WS, and WM rotation modes, respectively. The average CO2 emission rate was 4.69, 4.61, and 4.53 μmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively. The average soil CO2 emission rate during the 2012 summer growth period under the WF, WS, and WM rotation modes was 6.57, 4.73, and 7.01 μmol·m −2 ·s −1 , respectively, and it was 6.16, 5.94, and 6.20 μmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively.
The results of this study show that the three rotation modes produced significant differences in soil CO2 emission rates during the 2012-2013 winter wheat growing season, but there was no effect During the summer crop growing seasons ( Figure 5), the soil CO 2 emission differed significantly (p = 0.0095 and p = 0.03 in 2011 and 2012, respectively) between the three rotation modes. The average soil CO 2 emission rate of the WF rotation mode was the lowest of the three rotation modes, and the WM rotation was the highest. During the summer growth period of 2011, the average CO 2 emission rates were 2.75, 4.96, and 6.12 µmol·m −2 ·s −1 under the WF, WS, and WM rotation modes, respectively. The average CO 2 emission rate was 4.69, 4.61, and 4.53 µmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively. The average soil CO 2 emission rate during the 2012 summer growth period under the WF, WS, and WM rotation modes was 6.57, 4.73, and 7.01 µmol·m −2 ·s −1 , respectively, and it was 6.16, 5.94, and 6.20 µmol·m −2 ·s −1 under the TF, 0.8TF, and NF treatments, respectively.
The results of this study show that the three rotation modes produced significant differences in soil CO 2 emission rates during the 2012-2013 winter wheat growing season, but there was no effect of N fertilizer treatment ( Figure 5). The mean soil CO 2 emission rate under the WF, WS, and WM rotation modes was 9.39, 10.64, and 12.52 µmol·m −2 ·s −1 , respectively in the 2011-2012 winter wheat growing period, and 2.27, 1.29, and 1.70 µmol·m −2 ·s −1 , respectively, during the 2012-2013 winter wheat growing season. The mean soil CO 2 emission in the 2011-2012 winter wheat growing period under the TF, 0.8TF, and NF treatments was 10.21, 10.56, and 11.77 µmol·m −2 ·s −1 , respectively, and it was 1.63, 1.85, and 1.77 µmol·m −2 ·s −1 , respectively, during the 2012-2013 winter wheat growing period.

Soil Nutrients
The alkali-hydro N, available phosphorus, and available potassium concentrations varied under different crop rotations and N fertilizer treatments in the 0-40 cm soil layer over the two years ( Figure 6). No significant difference in alkali-hydro N concentration between the two years was found. The mean alkali-hydro N concentrations were 24 growing period, and 2.27, 1.29, and 1.70 μmol·m −2 ·s −1 , respectively, during the 2012-2013 winter wheat growing season. The mean soil CO2 emission in the 2011-2012 winter wheat growing period under the TF, 0.8TF, and NF treatments was 10.21, 10.56, and 11.77 μmol·m −2 ·s −1 , respectively, and it was 1.63, 1.85, and 1.77 μmol·m −2 ·s −1 , respectively, during the 2012-2013 winter wheat growing period.

Soil Nutrients
The alkali-hydro N, available phosphorus, and available potassium concentrations varied under different crop rotations and N fertilizer treatments in the 0-40 cm soil layer over the two years ( Figure  6). No significant difference in alkali-hydro N concentration between the two years was found. The mean alkali-hydro N concentrations were 24

Relationship between Soil Organic Carbon and CO 2 Emission and Treatment Factors
Crop and soil variability affected soil CO 2 emission. The relationship between soil temperature and soil CO 2 emissions was identified by a redundancy analysis (RDA) (Figure 7). Soil temperature was strongly correlated with soil CO 2 emission rate and SOC concentration during the two-year study Sustainability 2020, 12, 5271 9 of 14 under the three N fertilizer treatments and the three rotation modes, and explained 16.8, 47.40, and 24.10% of variation in the WF, WS, and WM rotations, respectively. Correlation analysis showed that soil CO 2 emission rate was negatively correlated with soil water content, and that soil water content may decrease the release of CO 2 from deep soil.

Relationship between Soil Organic Carbon and CO2 Emission and Treatment Factors
Crop and soil variability affected soil CO2 emission. The relationship between soil temperature and soil CO2 emissions was identified by a redundancy analysis (RDA) (Figure 7). Soil temperature was strongly correlated with soil CO2 emission rate and SOC concentration during the two-year study under the three N fertilizer treatments and the three rotation modes, and explained 16.8, 47.40, and 24.10% of variation in the WF, WS, and WM rotations, respectively. Correlation analysis showed that soil CO2 emission rate was negatively correlated with soil water content, and that soil water content may decrease the release of CO2 from deep soil.

Effect on Soil Temperature, Soil Moisture, SOC, and Soil CO2 Emission in the Rotation Mode
It has been reported that adopting intensified rotation can lower the carbon footprint in semiarid areas [34]. Winter wheat plus summer soybean or summer maize are the most common double cropping systems in the Ghuanzhong region, Shaanxi Province in north-central China [35]. A previous study found that the yields of crop residue varied under different rotation modes [36]. The quality and quantity of crop residues returned to the soil changes with the different cropping structures, with the C/N ratio of the residues varying with the different crop rotation modes [37], which can influence CO2 emissions, and concurs with previous findings [38]. The current study shows that the average CO2 emission from the WM and WF rotation modes was higher than that from the WS rotation during the growing seasons over the two-year study period. The soil CO2 emission rate was low during the summer crop growth period under the WF rotation mode, possibly due to the lack of plants during this period. The average soil CO2 emission rate decreased with increasing amounts of N fertilizer under the WF rotation mode, while the average soil CO2 emission rate under the WM rotation increased with increasing amounts of N fertilizer. The average soil CO2 emission rate was similar under the WS rotation among N fertilizer treatments. A crop rotation strategy that included a legume reduced carbon losses, which is in line with the findings of Miao [18].
Crop communities are drivers of soil CO2 emissions [39]. Soil CO2 emission rate had a stronger correlation with SOC under the WF rotation mode in the 0-10 cm soil layer than the correlation under the WS and WM rotations (Figure 7). The main reason for this is that the WF rotation did not include

Effect on Soil Temperature, Soil Moisture, SOC, and Soil CO 2 Emission in the Rotation Mode
It has been reported that adopting intensified rotation can lower the carbon footprint in semiarid areas [34]. Winter wheat plus summer soybean or summer maize are the most common double cropping systems in the Ghuanzhong region, Shaanxi Province in north-central China [35]. A previous study found that the yields of crop residue varied under different rotation modes [36]. The quality and quantity of crop residues returned to the soil changes with the different cropping structures, with the C/N ratio of the residues varying with the different crop rotation modes [37], which can influence CO 2 emissions, and concurs with previous findings [38]. The current study shows that the average CO 2 emission from the WM and WF rotation modes was higher than that from the WS rotation during the growing seasons over the two-year study period. The soil CO 2 emission rate was low during the summer crop growth period under the WF rotation mode, possibly due to the lack of plants during this period. The average soil CO 2 emission rate decreased with increasing amounts of N fertilizer under the WF rotation mode, while the average soil CO 2 emission rate under the WM rotation increased with increasing amounts of N fertilizer. The average soil CO 2 emission rate was similar under the WS rotation among N fertilizer treatments. A crop rotation strategy that included a legume reduced carbon losses, which is in line with the findings of Miao [18].
Crop communities are drivers of soil CO 2 emissions [39]. Soil CO 2 emission rate had a stronger correlation with SOC under the WF rotation mode in the 0-10 cm soil layer than the correlation under the WS and WM rotations (Figure 7). The main reason for this is that the WF rotation did not include crops during the summer period, contrary to the WS and WM crop rotation, leading to a decrease in crop root CO 2 emission. Soil CO 2 emission under the WF rotation came from the transformation of organic matter and microbial CO 2 emission, which is a source of CO 2 emissions [40]. Microbial activity is positively influenced by soil organic matter content and climate change [41]. Soil microbial activities are enhanced when legumes and non-legumes are grown together compared to non-legumes grown alone [42], and legume and non-legumes grown individually [43]. Soil CO 2 emissions are strongly affected by plant roots [4], which was reflected in the differences between the summer crop soils and the summer fallow soil in the current study.
Changes in land use might affect efforts to improve the quantity and activity of the SOC pool [44]. In the current study, SOC in the topsoil (0-10 cm) increased over the two-year study period under all three rotations. Early data have shown that proper adoption of crop rotation can improve soil chemical and physical properties, increase the quality of SOC and soil environmental conditions, limit soil erosion, and increase agricultural productivity, thus contributing to sustainable agriculture [11]. Higher SOC content under the WM rotation than WF and WS rotation indicates that soybean, a leguminous crop, decreased SOC content despite increasing soil N content, which concurs with previous studies [11]. Ardell (2012) also reported that the rate of gain in SOC under continuous corn rotations was higher than that under a corn-winter wheat-grain sorghum-soybean rotation [37]. This might be due to different crop rotations influencing the amount of crop root residue and root exudation that is returned to the soil, which in turn affects C and N content, and can be used to increase SOC [30]. However, soil organic matter changes slowly over time [11].
Temperature was the main factor driving variation in soil CO 2 emission rates [30]. This may be due to its effect on the decomposition of plant residues and SOC by microbes, the diffusion of enzymes, and root CO 2 emission [30,41]. A redundancy analysis (RDA) showed that soil temperature at 10 cm was positively correlated with soil CO 2 emission. The main reason for this might be that the temperature of soil at 5 cm was affected by air temperature. Soil temperature was measured between 9:00 and 11:00 a.m., and soil temperature at 5 cm increased with air temperature, while the response in soil temperature at 10 cm was delayed.
Soil moisture plays an integral role in determining CO 2 emissions; a prolonged period with deficient or excess water in the soil can cause soil CO 2 emission rates to fall [45]. The rate of soil CO 2 emission under winter wheat in 2011-2012 was higher than that in 2012-2013, which may be due to the low level of soil moisture in 2011-2012 compared with 2012-2013, and the soil temperature was higher in 2011-2012 than it was in 2012-2013. During the winter wheat growth period, precipitation was below average, and thus, the soil was almost dry with minimal water content. Soil water can restrict the diffusion of soil CO 2 in soil pores, and lower temperatures reduce the activity of microorganisms [6]. Redundancy analysis (RDA) showed that soil CO 2 emission was negatively correlated with soil moisture and positively correlated with soil temperature. However, soil water responded to precipitation events during the summer crop period with higher water levels. Water content was greater in the WF rotation than in the WS and WM rotations, which may be due to the absence of plants during the fallow period, which reduces the consumption of soil water, which is in line with the observations made in previous studies [38].

Influence of N Fertilizer on Soil Organic Carbon (SOC), Soil Temperature, Soil Moisture, and Soil CO 2 Emissions
China's N fertilizer production exceeds its consumption, and N fertilizer application rates exceed crop requirements for maximum yield, which is a common problem in the North China Plain [46]. Small-scale farmers lack knowledge of N management, and hand application of fertilizers (to increase yield) has led to excessive application and low fertilizer use efficiency in farming systems in China [18]. Improving fertilizer use efficiency in crop production is a key issue for addressing the triple problem of food security, environmental degradation, and climate change, and has major ramifications for global emissions [22,23]. Agricultural CO 2 emissions are augmented by N fertilization [19,47]. This is probably due to increased root CO 2 emission, large amounts of crop root residue, and root exudations, which increase carbon substrate availability, and lead to soil organic matter decomposition and enhanced microbial activity [47]. However, N fertilization can have variable effects on CO 2 emissions [48]. Compared to unfertilized soil, another study showed that N fertilization decreased CO 2 emissions by 27-42% [49]. In the current study, the soil CO 2 emission rates under different N fertilizer treatments were in the following order: NF > 0.8TF > TF. This indicates that N application can reduce CO 2 emissions from croplands. This is likely due to reduced soil organic matter mineralization [49].
Due to lack of fertilizer treatment, decomposition of native soil organic matter was likely to have occurred [19,48].
Plant yield, biomass, tiller number, and leaf size of wheat are affected by N fertilizer application and result in different soil temperature and water moisture levels [14]. Crop height, tiller height, and leaf area of winter wheat under N fertilizer application was significantly greater than it was with no N fertilizer application. In the current study, average temperature across the winter wheat growth periods was higher under the NF treatment than the 0.8TF and TF treatments. The temperatures of soil under NF treatment were higher than those under the 0.8TF and TF treatments, because the shading of straw in the NF treatment was lower than it was under 0.8TF and TF treatments, in agreement with Sainju et al. [38]. Crop production is limited by soil water content in dryland agriculture in China [50]. There were no significant differences in average soil moisture levels under the three N fertilizer treatments.
A previous study found that enhancing soil carbon sequestration can improve soil quality, increase net primary productivity of plant biomass, and reduce agriculture's contribution to soil CO 2 emissions [34]. The results demonstrated that the concentration of easily decomposed SOC is the main factor affecting the soil CO 2 emission rate, which is influenced by N fertilization. A similar conclusion was reached by Ding [48]. However, it is debatable whether N fertilizer application is beneficial for SOC sequestration. Limited N application may reduce total crop biomass [23], whereas optimized N application can improve plant growth and increase the amount of residues returned to the soil. We found that optimized N application can increase soil organic matter content and carbon sequestration [51]. In this study, SOC content in the 0-10 cm soil layer increased by 10.30, 20.12, and 20.17% between the summer of 2011 and the summer of 2013 under NF, 0.8TF, and TF, respectively. However, chemical fertilizer application can accelerate the decomposition of original SOC [47]. As a result, the average SOC content in the 0-10 cm soil layer of the 0.8TF treatment under all the three rotation modes was higher than that under the other two fertilizer treatments.

Conclusions
This research demonstrates that crop rotations and N fertilizer application can affect soil CO 2 emission and SOC in the field. The different crops in the three rotations under the three N fertilizer treatments led to different soil CO 2 emission rates, soil temperatures, and soil water moisture levels, with the soil CO 2 emission rate under the WM rotation higher than the WF and WS rotations. The mean soil CO 2 emission rate under the NF treatments was higher than those under the TF and 0.8TF treatments. Based on this, it is concluded that under the WS rotation the soil CO 2 emission rate was lower and the N use efficiency was higher because of the leguminous summer crop, which produces symbiotically fixed N. In the future, WS rotation with higher soil nutrient content, lower soil CO 2 emissions, and reduced fertilizer application will be the most effective protocol for sustainable agriculture, as well as being a better option in this region.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2071-1050/12/13/5271/ s1- Figure S1: Changes in soil temperature under different crop rotations and N fertilizer treatments. Bars show means± s. e. m. Figure S2: Changes in soil water content under different crop rotations and N fertilizer treatments. Figure S3: Soil organic carbon content of the 0-100 cm soil layer under three N fertilizer treatments and three rotation treatments. Figure S4: Changes in soil respiration rate under different crop rotations and fertilizer treatments.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.