Field experiments were conducted at the experimental farm at Yangtze University (30°36′ N, 112°08′ E, 34 m asl) in Jingzhou, Hubei Province, China in 2017–2018 and 2020–2021.

This experiment was conducted with two cropping rotation systems in both the cases, we only focused on the wheat within each rotation (Figure 1):

Dryland crop rotation system: The summer crops are dryland crops, mainly corn and soybeans, and winter wheat is DW.

Rice-wheat rotation system: The summer crop is rice and winter wheat is rice stubble wheat (RW). This cropping rotation system has a unique dry–wet cycle soil regime.

The daily mean temperature and rainfall during the wheat-growing season are shown in Figure 2. Meteorological data were collected during each growing season using an automatic weather station adjacent to experimental trials conducted here.

Before beginning the experiment, both the cropping systems had been used primarily for wheat production for more than 10 years. The tillage soil layer (0–20 cm) of the two cropping systems was as follows. Soil organic matter 11.00 g kg^–1, available nitrogen 82.03 mg kg^–1, available phosphorus 15.20 mg kg^–1, available potassium 51.11 mg kg^–1, and pH 7.8 in DW, and soil organic matter 12.37 g kg^–1, available nitrogen 51.22 mg kg^–1, available phosphorus 12.07 mg kg^–1, available potassium 52.74 mg kg^–1, and pH 7.79 in RW.

The treatments were arranged in a split-plot design with a cropping system as the main plots and N treatment as the subplots. Nitrogen fertilization treatments (timing and amount) were designed in line with local management. The four N treatments were applied as follows: 0 kg N ha^–1 (N₀), 135 kg N ha^–1 (N₁₃₅), 180 kg N ha^–1 (N₁₈₀), and 225 kg N ha^–1 (N₂₂₅), of which 180 kg N ha^–1 is the local conventional N application rate. N fertilizer was applied at the basal, tillering, and jointing initiation stages in a ratio of 1: 1: 1, being 45 kg N ha^–1 for the N₁₃₅ treatment, 60 kg N ha^–1 for the N₁₈₀ treatment, and 75 kg N ha^–1 for the N₂₂₅ treatment, respectively. Each plot is repeated four times, arranged in random block groups, with a plot area of 12 m² (2 m × 6 m). The experimental design and wheat growth process are shown in Figure 1. In the experiment, “Zhengmai 9023” was used as the tested variety, and the planting density was 2.25 million plants ha^–1. N fertilizer was applied in the form of urea (46% N) that was ploughed into the soil during sowing and was spread as top dressing. Plots were supplied with P (105 kg P₂O₅ ha^–1, calcium superphosphate) and K (105 kg K₂O ha^–1 potassium sulfate) fertilizer during the sowing period. We did not need to irrigate during the two growing seasons due to the abundant rainfall. Herbicides, pesticides, and fungicides were sprayed according to standard growing practices to avoid yield loss.

We selected mature plants from a 2 m² harvest area in the middle of each plot to determine the yield and yield components. The spike number was enumerated in each plot and recorded prior to harvest. All spikes were manually harvested, threshed, and weighed to calculate grain yield. The average number of kernels per spike was enumerated from 30 spikes. One thousand random kernels from each harvested grain were weighed to calculate a 1,000-kernel weight. Grain moisture was measured using a grain analyzer (InfratecTM, Foss, Denmark). Grain yield and 1,000-kernel weight were adjusted to 13% moisture.

The shoot biomass was harvested from twenty mature plants in each plot at the flowering and maturity. Samples were oven-dried at 65°C for at least 48 h until they were a constant weight. Then, pulverized with a plant grinder, followed by digestion in sulfuric acid at 400°C. Then, the N contents in the plant were measured with an automatic Kjeldahl analyzer (Kjeltec-8400, FOSS, Denmark) and the values were used to calculate the agronomic efficiency of nitrogen (AE_N) and physiological efficiency of nitrogen (PE_N).

AE_N = (Grain yield in the plot that received N applied–grain yield in the zero-N control)/(amount of N applied).

Which represents the amount of yield increased per unit N fertilizer, reflecting the economic performance of N fertilizer in terms of grain production.

PE_N = (Grain yield in the plot that received N applied–grain yield in the zero-N control)/(total aboveground plant N accumulation in the plot received N applied–total aboveground plant N accumulation in the zero-N control).

Which represents the yield increase per kg increase in N uptake from fertilizer. It expresses the efficiency with which the plant utilizes N taken up to produce fruit dry biomass.

Soil samples were collected from each plot at the flowering (March 28, 2021), midfilling (April 14, 2021), and maturity (May 8, 2021) of wheat. Soil samples to a depth of 60 cm were obtained in triplicate by a soil auger (5-cm diameter) at 20-cm intervals. In total, 10 g of each sample was taken for determinations of nitrate (NO₃^–) and ammonium (NH₄⁺), and the remaining dose was used for a gravimetric water content determination. The 10 g of the soil sample was extracted with 50 ml of 1 mol l^–1 KCl to obtain the nitrate and ammonium concentrations in the soil. All nitrate and ammonium concentrations in various solutions were determined using A SMARTCHEN automatic intermittent chemical analyzer. The gravimetric soil moisture content of the samples was subsequently measured using the oven-drying method (105°C for 24 h).

Data were analyzed using a three-way (year, cropping system, and nitrogen management) ANOVA with SAS 9.2 (SAS Institute, Cary, NC, United States). Treatment means were compared using the least significant differences (LSD) at P < 0.05.

The daily mean temperatures for the wheat filling period were 20.4°C in 2017–2018 and 17.6°C in 2020–2021. The reduced temperature in the post-flowering period extends the length of the grain filling period in 2020–2021(Figure 2). The total rainfall for the wheat growing season was 351.1 mm in 2017–2018 and 263.8 mm in 2020–2021. In 2017–2018, the rainfall from wintering to jointing of wheat was 149.4 mm, accounting for 43% of the total rainfall during the growth period. In 2020–2021, the rainfall from the winter to the jointing of wheat was only 44.5 mm. The rainfall in this year was mainly concentrated in the filling period reached 132.6 mm, accounting for 50% of the rainfall in the entire growth period of wheat, 39% more than that in the same period in 2017–2018. Compared with 2017–2018, the frequency of rainfall in the filling period in 2020–2021 is higher. Rainy days accounted for 44% of the total filling days. The daily mean rainfall of the rainy days in the filling period was 7.0 mm, of which 57% was concentrated in the early stage of filling.

Figure 3 shows the 0–60 cm soil moisture content during the grain filling period of DW and rice stubble wheat (RW) in 2020–2021. The soil moisture content of RW was significantly (P < 0.05) higher than DW (except for 0–20 cm at maturity), higher 0.6–3.2% at 0–20 cm, 2.2–4.1% at 20–40 cm, and 2.1–6.9% at 40–60 cm soil depth than DW.

Increasing N significantly improved grain yield in DW and RW, with significant differences between years (Table 1). Compared with N0, the yield of N supply increased by 1.2–1.4 times (RW), 0.5–0.7 times (DW) in 2017–2018; the yield increased by 2.0–2.5 times (RW), 1.6–1.7 times (DW) in 2020–2021, increased N supply has a greater effect on the increase in RW. The yield of DW was higher than that of RW, which was more pronounced in the N₀ and N₁₃₅ treatment in 2020–2021. When the N supply was increased from 135 to 180 kg ha^–1, the yield increase of DW and RW will only reach significant levels in 2020–2021. When the N supply was increased from 180 to 225 kg ha^–1, the yield increase of DW and RW was not significant in 2017–2018 and 2020–2021.

The yield components of wheat were both significantly (P < 0.05) affected by planting patterns and N treatments (Table 1). The spike number of DW was 4–26.3% higher than that of RW under the same treatment, while RW has an advantage in the kernels per spike. Increased N supply has a difference in the increase of the spike number and kernels per spike of DW and RW. The highest increase in the spike number of DW and RW by increasing N supply by an average of 65.3 and 63.8%, respectively, over the two wheat growing seasons. The highest increase in the kernels per spike of DW was 28.3% on average, while RW could reach 87.8%. The results showed that increased N supply has a greater effect on increasing the kernels per spike in RW. Affected by the low rainfall from the tillering to the jointing of wheat and the excessive rainfall during the wheat filling in 2020–2021 (Figure 2), the spike number and kernels per spike of wheat under the same treatment in 2020–2021 was significantly lower than that in 2017–2018, and the 1,000-grain weight was not significant differences.

There were significant differences in the N accumulation and NUE between DW and RW (Table 2). Under the same N supply, the N accumulation in plant tissue and kernel dry matter of DW in 2017–2018 was 9.2–168.9% and 31.2–199.4% higher than at RW. There were significant differences only under N₀ treatment in 2020–2021. The agronomic efficiency of nitrogen (AE_N) and physiological efficiency of nitrogen (PE_N) at RW were significantly higher than at DW by 11.4–17.4 and 15.3–25.6 kg kg^–1. RW was also higher than DW in 2020–2021, but the difference was not significant. The plant N accumulation and NUE of DW and RW were different in years. The N accumulation in 2017–2018 was higher than that in 2020–2021, and there was a significant difference between DW and RW in the first year. The second year also showed the same regular difference, but the difference did not reach a significant level.

Increased N supply has a significant impact on the plant N accumulation and NUE at DW and RW (Table 2). With the increase of N supply, the N accumulation in plant tissue and kernel dry matter of DW at flowering and maturity increased significantly. Compared with N₁₃₅ treatment, the N accumulation at flowering and maturity of N₁₈₀ treatment increased by an average of 26.6 and 23.9%, respectively, over the two wheat growing seasons, and RW increased by 33.3 and 15.9%. The corresponding N accumulation in the kernel of DW and RW increased by 20.2 and 16.4%, respectively. The increase of N₂₂₅ treatment compared with N₁₈₀ treatment reached a significant level only in DW in 2017–2018 and RW in 2020–2021. With the increase of N supply, AE_N and PE_N of DW decreased significantly, but AE_N and PE_N of RW had no significant difference between nitrogen treatment (except for AE_N in 2017–2018).

In the flowering, both DW and RW 0–20 cm soil NO₃^–-N increased with the increase of N supply, and both reached the maximum under N₂₂₅ treatment (Figure 4). Which were significantly increased by 4.3–7.3 mg kg^–1 at DW, 3.6–4.9 mg kg^–1 at RW, compared with other treatments; But at 20–40 cm soil layer, the increase of N supply gradually reduced the soil NO₃^––N content. The DW soil NO₃^––N content of N₂₂₅ treatment increased 2.4–3.6 mg kg^–1 compared with other n treatments, while RW Only increased by 0.4–0.5 mg kg^–1. With the deepening of the soil layer, the difference in the soil NO₃^––N content at DW and RW between N treatments gradually decreased. In the middle of filling, only the 0–20 cm soil NO₃^––N content at DW has significant differences between N treatments, while at RW was not significant; There was no significant difference in the NO₃^––N content at DW in 20–60 cm soil between N treatments, but the N₂₂₅ treatment significantly increased 0.8–2.7 mg kg^–1 compared with other N treatments in RW. In the maturity, the 0–40 cm soil NO₃^––N content at DW and RW had no significant difference between N treatments.

The content of NH₄⁺–N in the soil is low, only 7.4–46.2% of the corresponding nitrate-nitrogen content (Figure 4). In the flowering, increased N supply increased the 0–20 cm soil NH₄⁺–N content at DW and RW. Compared with other N treatments, N₂₂₅ treatment increased by 0.02–0.45 mg kg^–1 under DW and RW. The NH₄⁺–N in 20–60 cm soil was not significantly different between N treatments. In the middle of filling, there were significant differences in the NH₄⁺–N at DW in 0–60 cm soil between N treatments, while the difference in RW was only in the 0–40 cm soil layer. At maturity, there were no significant differences in NH₄⁺–N between DW and RW in the 0–20 cm soil layer, but there were significant differences between the 40 and 60 cm soil layer.

The type of soil texture and the depth of the soil layer are the main factors that affect the soil moisture pattern. From the perspective of the vertical change characteristics of soil water content, the water content of the 0–60 cm soil layer increases with the increase of depth (Figure 3). Because there are more residual straws in the soil surface layer, the decomposed humus promotes the formation of soil porosity (Turmel et al., 2015), and the wheat root system is mainly distributed in the surface layer (Yang et al., 2021). The soil surface layer has a large infiltration rate, which is not conducive to the preservation of soil moisture. Therefore, the soil moisture content of the soil surface layer is likely to fluctuate with rainfall. There was a big difference in soil moisture between DW and RW. Rotation methods will change the physical properties of the soil. Long-term paddy-rice rotation promotes soil granulation. The soil texture of RW was stickier and its soil moisture content is 2.2–4.3% higher than that of DW on average (Figure 3). This may be mainly because the rice stubble wheat soil with fine texture and high field water holding capacity will more effectively store rainwater with high rainfall (Yoo et al., 1998).

The concentration of mineral nitrogen is an important index to evaluate the available nitrogen supply capacity of the soil, and the change of soil moisture is one of the most important environmental factors affecting nitrogen mineralization and nitrification. Studies have suggested that nitrification is more sensitive to water than ammonification, and the nitrification rate will be inhibited when the soil moisture is high (Schimel and Parton, 1986; Wang et al., 2021). This is mainly because the increase in rainfall results in a large soil moisture content, which is not conducive to the diffusion of oxygen in the soil. Nitrifying bacteria prefer aerobic conditions. Excessive rainfall makes clay loam soils with higher field water holding capacity more likely to form an anaerobic environment, inhibiting nitrifying bacteria and nitrification (Cameron et al., 2013). Soil available nitrogen supply was affected by soil water content, and rice stubble wheat has a greater decrease in nitrate-nitrogen content in the middle of grain filling when the soil water content was higher. Armstrong et al. (1996) found that the response of dryland sorghum to nitrogen application largely depends on the soil moisture content, and the response to nitrogen increase depends on the time of rainfall. The results of this study can also prove this point. Due to the influence of soil water content and high rainfall frequency, the nitrate-nitrogen content of the surface soil of rice stubble wheat is less responsive to increased nitrogen fertilizer application than DW. This is because of the three ways of soil nutrient loss (gas loss, leaching loss and runoff, and erosion loss), both leaching loss and runoff loss are highly susceptible to rainfall intensity. The Jianghan Plain has abundant rainfall, with a total rainfall of 132.6 mm during the wheat filling period in 2020–2021, which is a humid climate zone (Figure 2). The soil nitrogen loss pathways of rice stubble wheat and DW have a great relationship with soil texture and soil moisture. When the rainfall of the fine-textured paddy soil exceeds the infiltration amount of the soil, the field water holding capacity will reach a saturated state, and the field water is more likely to form surface runoff (Zhao et al., 2012); The rough soil has strong water conductivity, fast soil infiltration rate, and more serious leaching loss (Sugita and Nakane, 2007). We suggest that the interplay between soil water and soil nitrogen in crop rotations warrants further attention: part of this could include direct quantification of the association between soil moisture and soil nitrogen at the beginning of the cropping phase and subsequent implications for both growth and NUE.

This study adds significant novelty to the scientific literature as we compare: (1) cropping systems rotational effects (RW or soybean/corn–wheat) on soil nitrogen and crop NUE and (2) our contrast between RW and DW systems allows direct comparison of the effects of residual soil moisture on subsequent crop growth and NUE. Indeed, soil mineral nitrogen concentration is often linked with the vegetative growth of wheat. Some studies have shown that high levels of mineral N (NO₃^– and NH₄⁺) stimulate shoot biomass accumulation of fertilized treatments, neither seed yield nor seed nitrogen concentration was affected by soil mineral N availability (Voisin et al., 2002). In contrast, this study has shown that the yield and nitrogen accumulation of RW were related to nitrogen application rate and soil mineral nitrogen content at flowering (Table 2). Increasing nitrogen application significantly increased the soil mineral nitrogen content of rice stubble and wheat, and increased plant nitrogen accumulation. Compared to the RW, plant N accumulation at maturity was greater at the DW. Only when the N accumulation at DW at the flowering stage was greater than at RW, the N accumulation of at DW at maturity was greater at RW (Table 2).

Excessive rainfall at the time of fertilization may cause loss of external nitrogen sources at RW soil because the amount of fertilizer-N applied before flowering initiation was similar at the two systems. Therefore, in areas with high rainfall, RW may be more suitable for small and multiple fertilization, especially to avoid the high-frequency rainfall period (March–May). Even though Ney et al. (1997) stressed that any diagnosis using the critical nitrogen concentration was best done before seed filling. However, the more rainfall from March to May in Jianghan Plain makes the nitrogen supply in different wheat growing environments different, so we believe that it is equally important to understand the soil nitrogen supply in the later stages of wheat growth.

Under low nitrogen (0–135 kg ha^–1), the yield of DW was higher than that of RW, and under conditions of sufficient nitrogen supply, RW can also obtain the same yield as DW. The results show that the soil of DW has a higher basic nitrogen supply capacity. In this experiment, DW has a higher spike number and nitrogen accumulation at flowering. The soil of DW provides more nitrogen in the early growth period and has more tillers at wintering and spike number at harvest (Yin et al., 2018). This may be the reason for the higher yield of DW and the higher accumulation of nitrogen at flowering. Compared with DW, the conversion of paddy fields to dryland causes changes in the physical and chemical properties of the soil. Poor soil aeration has an adverse effect on the emergence of RW and the growth of the seedling stage (Nawaz et al., 2016). Therefore, the spike number at maturity of RW is less than that of DW. The yield increase effect of RW was mainly that the number of kernels per spike was increased by 71.2–84.4% (Table 1). Due to decreased shoots per unit area and alleviated competition for resources between shoots (Guo and Schnurbusch, 2015).

Prior to this study, the greater root length density is an important root trait for the crop with high grain yield and higher NUE under N-limited conditions or at lower N rates (Ju et al., 2015). Our previous results indicated that the root length density at DW was greater at RW, especially in deep soil layers. This may also be the reason why the N accumulation at DW was greater at RW (Yang et al., 2021). AE_N refers to the amount of yield increased per unit N fertilizer, reflecting the economic performance of N fertilizer in terms of grain production. Although the N accumulation at RW was lower than at DW, it can reach a similar yield level at DW. Therefore, RW has greater N use efficiency. Higher nitrogen accumulation level or NUE can achieve high yield.