Enhancing Maize Yield and Resource Efﬁciency through Controlled-Release Nitrogen Fertilization on the Semiarid Loess Plateau

: Drought stress is one of the premier limitations to global agricultural production. Increasing water and nitrogen (N) use efﬁciencies in dryland agroecosystems to maintain high agricultural output are key responsibilities to assure food security, especially on the semiarid Loess Plateau region of China, as it is one of the important grain production areas in China. The impact of controlled-release urea (CRU) on the soil water content, soil enzyme activities, soil N content, biomass accumulation, grain yield, water use efﬁciency (WUE), and agronomic use efﬁciency of N fertilizer (AEN) were examined on the maize production of the rainfed Loess Plateau during 2020–2021. Two-growing-season ﬁeld treatments at the Zhengyuan Agri-ecological Station, Qingyang, Gansu, including six N treatments, were investigated for maize: a control without N fertilization (CK) and ﬁve application proportions of CRU (i.e., 0, 30, 50, 70, and 100%CRU) under a N rate of 225 kg ha − 1 . Results showed that compared with common urea (0%CRU), on average, CRU applications signiﬁcantly increased soil enzyme activity related to N conversion and improved biomass accumulation by 4–11% at the silking stage and by 2–12% at the maturity stage, respectively. As the proportion of CRU increased, the grain no. per ear, 100-grain weight, and harvest index ﬁrst increased and then decreased. Grain yield was increased by 5.3, 11.4, 20.1, and 5.7% under 30, 50, 70 and 100%CRU, respectively, compared to common urea. Compared to common urea, 70%CRU combined with 30% common urea achieved the highest yield. These results indicate that optimal controlled-release N fertilization increases the yield and water and nitrogen use efﬁciencies of maize, and 70%CRU combined with 30% common urea under a single application of nitrogen fertilizer at sowing was the optimal application proportion of controlled-release urea for increasing water and nitrogen use efﬁciencies in dryland agroecosystems. The results of this study can provide a theoretical basis for the efﬁcient fertilization of maize on the semiarid Loess Plateau of China


Introduction
Water and nitrogen (N) availability limit crop productivity globally more than most other environmental factors. Enhancing crop productivity and optimizing water and nutrient utilization are crucial worldwide to guarantee food security and enhance ecological conditions, particularly in arid regions [1][2][3][4]. Nevertheless, China possesses a substantial expanse of arid land in its northern territories, constituting nearly 56 percent of the country's entire land mass in the region [1,5,6], and dry farming in this region accounts for about 50% of the total agricultural production area. Therefore, dryland agriculture is an important mode of agricultural production in China [1], and the grain-yield-increasing potential of China will be in the dryland regions in the future. The Loess Plateau, spread over approximately 64 million hectares, supports about 100 million people and is the major dryland farming area of China [7]. Limited rainfall, high evaporation, severe soil loss, and low soil fertility constrain agricultural development in this region. Efficiently achieving high yields in dryland agriculture holds great significance for securing food supplies and the ecological stability of the Loess Plateau. Numerous agricultural management techniques aimed at improving maize production in dryland regions have been extensively explored in recent decades, among which terrace construction, fertilizer application, and the use of plastic film mulch have emerged as the most successful approaches. Maize has become a dominant crop after the application of plastic film covering technology on the semiarid Loess Plateau because the projected yield potential of maize is much higher than other crops, such as wheat (Triticum aestivum L.) and potato (Solanum tuberosum L.) [8].
The utilization of nitrogen (N) fertilization is a significant approach to enhance crop production [4]. It has been demonstrated to increase crop yields by 30-50%, and more than half of the global population relies on such fertilizers to sustain food production [9,10]. However, since the availability of N to crops is influenced by soil moisture, drought conditions can lead to simultaneous limitations in both water and nitrogen for crops. Consequently, in China, where achieving high grain yields is a priority, excessive N fertilizer application has become prevalent, particularly in dryland areas. Unfortunately, this disproportionate use of N fertilizers diminishes resource utilization efficiency and has adverse effects on the environment. For example, crop N recovery efficiency in China averages 32% compared to the world average of 55% while resulting in a negative effect on the environment, such as greenhouse gas emissions [11,12], ammonia volatilization [13,14], and water pollution [15]. Therefore, there is an urgent need for improved nitrogen management practices that optimize crop yield, decrease environmental influence, and ensure soil productivity to address food security concerns [11].
Common urea (solid granular urea) has been the most heavily applied N fertilizer in the maize production in China [16][17][18][19]. However, since the multiple topdressing of fertilizers consumes additional investments of labor, time, and equipment [16], it also promotes greenhouse gas emissions, especially ammonia volatilization (NH 3 ) and nitrous oxide emissions [17]. The utilization of enhanced-efficiency fertilizers, such as urease inhibitors, nitrification inhibitors, and slow/controlled-release urea, has been suggested as a central strategy for reducing nitrogen (N) losses [20]. Controlled-release urea (CRU) fertilizer application offers a result to enhance N utilization, decrease economic expenses, and boost crop yields. CRU is specifically designed to discharge N at a controlled pace that aligns with crop demand, thereby mitigating the contamination of the environment while sustaining crop productivity [16,21]. Mixed controlled-release and common urea may be synchronized with the N requirements of high-yielding maize [22]. Several studies have indicated that CRU significantly enhances maize yield and N utilization efficiency when compared to customary urea [23][24][25][26]. Guo et al. (2017) demonstrated that a mixture of CRU and urea at 1:2 and 240 kg N ha −1 achieved the highest economic return [27]. Bai et al. (2021) found that the one-time application of a mixture of controlled-release and solid granular urea (1:2) reduced the carbon footprint and increased grain yield [16].
However, in maize production in the arid and semiarid areas of northwestern China, information related to one-time fertilization with combined common urea and CRU is still very limited. Contrasting outcomes have also been reported by other studies [28][29][30]. These conflicting results could potentially be attributed to variations in experimental locations, N fertilizer application rates, and/or the proportion of CRU application. However, the majority of researchers have primarily focused on examining the impact of N fertilizer on crop yield [24,31], with only a small number of authors giving attention to determining the Agronomy 2023, 13, 2320 3 of 17 optimal application rate of CRU for maximizing farm profitability. The varying responses of maize yield to CRU highlight the need for further investigation into the specific conditions and methods for applying CRU. It is crucial to study the effects of combining the application of CRU and conventional urea on soil N management and the maize production system. To address this, a field study was conducted from 2020 to 2021 on the semiarid Loess Plateau to compare the continuous application of CRU with conventional urea. It was hypothesized that (i) CRU application to maize increases soil enzyme activities, increasing maize N uptake and biomass accumulation, and (ii) the optimal CRU application increases the yield and water and nitrogen use efficiencies of maize without increasing ET on the semiarid Loess Plateau. The main objectives of this study were to determine the effects of different combinations of CRU and urea on maize grain yield, water utilization, and water and N use efficiency in the natural field.

Site Description
The experiment in the field was carried out at the Northwest Dry Farming Nutrition and Fertilization Scientific Observation and Experiment Station of the Ministry of Agriculture and Rural Affairs (35 • 29 N, 107 • 29 E and elevation 1279 m) in Zhenyuan County, Qingyang City, Gansu Province, China, from April 2020 to October 2021. The site soil is classified as Cumuli-Ustic Isohumosols, according to Chinese soil taxonomy [32], which is a Calcaric Cambisol. The experimental site experiences an average annual precipitation of 540 mm, with more than half of it occurring between July and September. Rainfall in this region fluctuates greatly, and the probability of rainfall greater than 500 mm is only 57.1%. The annual water-surface evaporation is 1532 mm. The average annual temperature is 8.3 • C, and the average frost-free period is 170 d. The precipitation throughout the growing season is illustrated in Figure 1. Soil samples collected from the 0-20 cm depth at the experimental site in April 2020 exhibited a bulk density of 1.30 g cm −3 , a field capacity of 0.289 cm 3 cm −3 , and a wilting point of 0.098 cm 3 cm −3 . The pH of the soil, measured in a suspension of soil in 0.01 M CaCl 2 , was found to be 8.4 in the 0-20 cm soil layer. Furthermore, the soil analysis revealed the following nutrient values: total N content of 0.95 g kg −1 , soil organic matter of 12.1 g kg −1 , available phosphorus content of 15.1 mg kg −1 , and available potassium content of 145.7 mg kg −1 . on crop yield [24,31], with only a small number of authors giving attention to determining the optimal application rate of CRU for maximizing farm profitability. The varying responses of maize yield to CRU highlight the need for further investigation into the specific conditions and methods for applying CRU. It is crucial to study the effects of combining the application of CRU and conventional urea on soil N management and the maize production system. To address this, a field study was conducted from 2020 to 2021 on the semiarid Loess Plateau to compare the continuous application of CRU with conventional urea. It was hypothesized that (i) CRU application to maize increases soil enzyme activities, increasing maize N uptake and biomass accumulation, and (ii) the optimal CRU application increases the yield and water and nitrogen use efficiencies of maize without increasing ET on the semiarid Loess Plateau. The main objectives of this study were to determine the effects of different combinations of CRU and urea on maize grain yield, water utilization, and water and N use efficiency in the natural field.

Site Description
The experiment in the field was carried out at the Northwest Dry Farming Nutrition and Fertilization Scientific Observation and Experiment Station of the Ministry of Agriculture and Rural Affairs (35°29′ N, 107°29′ E and elevation 1279 m) in Zhenyuan County, Qingyang City, Gansu Province, China, from April 2020 to October 2021. The site soil is classified as Cumuli-Ustic Isohumosols, according to Chinese soil taxonomy [32], which is a Calcaric Cambisol. The experimental site experiences an average annual precipitation of 540 mm, with more than half of it occurring between July and September. Rainfall in this region fluctuates greatly, and the probability of rainfall greater than 500 mm is only 57.1%. The annual water-surface evaporation is 1532 mm. The average annual temperature is 8.3 °C, and the average frost-free period is 170 d. The precipitation throughout the growing season is illustrated in Figure 1. Soil samples collected from the 0-20 cm depth at the experimental site in April 2020 exhibited a bulk density of 1.30 g cm −3 , a field capacity of 0.289 cm 3 cm −3 , and a wilting point of 0.098 cm 3 cm −3 . The pH of the soil, measured in a suspension of soil in 0.01 M CaCl2, was found to be 8.4 in the 0-20 cm soil layer. Furthermore, the soil analysis revealed the following nutrient values: total N content of 0.95 g kg −1 , soil organic matter of 12.1 g kg −1 , available phosphorus content of 15.1 mg kg −1 , and available potassium content of 145.7 mg kg −1 .

Experimental Design
The experiment followed a randomized complete block design with three replications and focused on six treatments for maize. These treatments included a control group without N fertilization (CK) and five different ratios of controlled-release urea (CRU) application (0%, 30%, 50%, 70%, and 100%) at a fixed N rate of 225 kg ha −1 . The remaining N required for each treatment was supplemented with regular urea, as indicated in Table 1. The plot size was 22 m 2 (4.4 × 5 m), and the fertilization method and planting density adhered to the local practices for maize production in the region. All N fertilizers were applied at once as base fertilizers and were manually distributed over the soil surface prior to sowing and then plowed into the soil [17], and the N treatments remained consistent in the same plots each year. Polyethylene was used as a coating material for the production of CRU. Additionally, each year, all treatments received a basal application of 120 kg ha −1 P 2 O 5 in the form of triple superphosphate (16% P 2 O 5 ). As per standard farming practices in the region, we did not apply sulfur fertilizer in our experiment, and no potassium fertilizer was used in this experiment due to the apparently abundant potassium levels in the soils of the area [17][18][19]. In late April, all fertilizers were incorporated into the soil at a depth of 10 cm using a rotary tiller. Simultaneously, soil ridging and the placement of a transparent polyethylene film, measuring 120 cm in width and 0.01 mm in thickness, were carried out. The purpose of the ridges was to redirect raindrops toward the furrows where the maize crop was sown. Maize seeds of the Funong 821 variety were manually drilled into the furrows in late April, with a line spacing of 55 cm. The plant spacing within each row was 22.5 cm, resulting in a density of 75,000 plants per hectare. The maize crop was harvested in early October, and all maize residues were removed from the plots during harvest. The plastic film used for covering the soil was also removed using a residual film recycling machine. Subsequently, in late October, all plots underwent moldboard ploughing to a depth of 25 cm.

Sampling and Measurements
Daily rainfall data were collected from an automated weather station situated at the experimental site. Rainfall measurements were recorded and categorized as annual, fallow, and growing season rainfall, aligning with the different stages of maize growth in the specific region (as outlined in Table 2). To evaluate the fluctuations and conditions of annual rainfall, fallow period rainfall, and seasonal rainfall between the two growing seasons, a drought index (DI) based on rainfall was utilized. The calculation of the DI followed data collected in previous studies [4,18].
Each year, soil samples were collected from the surface down to a depth of 200 cm, with intervals of 20 cm per soil layer, at different stages: before sowing, during jointing, silking, and after the final harvest (GS92). These samples were taken to measure the soil water content. The fresh weight of the samples was promptly measured, and then the samples were dried in a forced-air oven at 105 • C until a constant weight was reached to determine the dry weight. The soil moisture was calculated based on the dry weight and the initial fresh weight of the samples. The soil water storage (SWS) was determined according to the following equations [4,18]: where SWS, SD, R, and Wm are soil water storage (mm), soil depth (m), soil bulk density (g cm −3 ), and soil moisture at different depths (%), respectively. Evapotranspiration (ET) is calculated by the soil water balance equation [4,7,18]: where SWCs is the soil water content at maize planting (mm), Rg is the seasonal rainfall (mm), and SWCh is the soil water content at harvest (mm). During the 2020-2021 growing seasons, at the silking stage, three random soil cores (with a diameter of 4.5 cm and height of 20 cm) were extracted from each plot within the 0-20 cm soil layer. These soil cores were collected for the purpose of measuring soil nutrient contents. The soil samples were air-dried, ground, and sieved through a 2.0 mm mesh. For analysis, 5 g of the dried samples was mixed with 50 mL of 2 M KCl solution, shaken for 60 min, and then assessed using an AA3 Automatic Continuous Flow Analyzer (Skalar, Breda, the Netherlands) to determine the levels of nitrate nitrogen (NO 3 − -N) and ammonium nitrogen (NH 4 + -N). To release and convert alkali-hydrolyzable N to NH 3 , 1.07 M NaOH and FeSO 4 powder was employed at a temperature of 40 • C for 24 h. The released NH 3 was subsequently absorbed using 2% (w/v) H 3 BO 3 solution and titrated with 0.005 M H 2 SO 4 .
Soil urease, nitrate reductase, nitrite reductase, sucrose, and catalase are crucial factors in the transformation and availability of nitrogen and phosphorus in the soil for crop uptake. Hence, these five enzymes were specifically chosen for this study to assess the impact of different N treatments on soil fertility. During the silking stage, three fresh soil samples were collected from each plot within the 0-20 cm soil layers to determine the activity of soil enzymes. The activity of soil urease was measured using the indophenol blue method, with the urease activity expressed as the amount of NH 4 + -N released per gram of fresh soil over a 24 h period at 37 • C, as outlined in the study by Ge et al. (2010). The catalase activity (0.1 mol L −1 KMnO 4 g −1 soil d −1 ) was measured by back-titrating residual H 2 O 2 with KMnO 4 [33]. Soil sucrase activity was measured by the 3,5-dinitrosalicylic acid methods [34]. Soil nitrate reductase and nitrite reductase activities, expressed as µg NO 2 − -N released g −1 fresh soil d −1 at 37 • C, were measured according to previous studies [35,36]. All determinations of enzymatic activities were performed in triplicates, and all values reported are their averages.
Plant dry mass (DM) was determined by randomly selecting one plant for each plot at different growth stages, including the seedling stage, jointing stage, tasseling stage, and milk ripening stage, respectively. Sample plants were removed directly above the nodal roots, and samples were oven-dried at 80 • C to a constant weight. At the time of physiological maturity, all maize plants within each plot were manually harvested from a designated area of 10 m 2 (2 m × 5 m). After the harvest, the aboveground dry mass was estimated, and the weight of air-dried grain was recorded for each plot to calculate the grain yield. Measurements of kernels per ear and 100-kernel weight were conducted during both the 2020 and 2021 growing seasons. The following parameters, related to dry matter accumulation and remobilization, were calculated following previous studies [37,38]. Harvest index (HI) and WUE were calculated as follows: where GY is the grain yield (kg ha −1 ), BY is the biomass yield (kg ha −1 ), WUE b is the water use efficiency for biomass (kg mm −1 ha −1 ), and WUEg is the water use efficiency for grain yield (kg mm −1 ha −1 ). The agronomic use efficiency of N fertilizer (AEN) refers to the augmentation in maize grain yield per unit of fertilizer nitrogen utilized which was calculated using the following formula: where Yn and Y0 are the grain yields in the fertilized plot and in the non-N-applicated plot (kg ha −1 ), respectively, and Nx is the amount of N fertilizer applied (kg N ha −1 ). The apparent nitrogen recovery (%) can be calculated as follows: where TNn and TN0 are the total N uptake by plants in the fertilized plot and in the non-N-applicated plot (kg ha −1 ), respectively, and Nx is the amount of N fertilizer applied (kg N ha −1 ). The physiological use efficiency of nitrogen (PhEN) can be calculated as follows: where Yn and Y0 are the grain yields in the fertilized plot and in the non-N-applicated plot (kg ha −1 ), respectively, and TNn and TN0 are the total N uptake by plants in the fertilized plot and in the non-N-applicated plot (kg ha −1 ), respectively.

Data Analysis
The statistical analysis was performed utilizing SAS software (SAS Institute, Cary, NC, USA, 2010) for conducting standard ANOVA and Duncan's tests (p ≤ 0.05). An analysis of variance was carried out for all variables with dependence in the study. The treatment effects were treated as fixed factors, while the growing season, replication, and all interactions with replicates were measured as random factors.

Weather Condition
Between the two years (Table 2), there was limited fluctuation in the amount of annual precipitation, with the majority occurring from July to September (Figure 1). The distribution of rainfall during the fallow period and cropping season differed across the two years. The fallow period experienced rainfall that closely resembled the long-term average of 125 mm, but it was higher in 2021 compared to the average. The seasonal rainfall Agronomy 2023, 13, 2320 7 of 17 did not align with the annual precipitation pattern. Although the growing season rainfall in 2021 was similar to the average of 405.5 mm, it exceeded the long-term average by 44.7 mm in 2020, indicating a wetter condition.

Effect of Controlled-Release N Fertilization on Soil Water Content and Evapotranspiration
In the 2020 growing season, soil water content was not significantly different at all growth stages among the N treatments (Figure 2A-D). In the 2021 growing season, soil water content under the treatment without N (CK) was greater in the 0-60 cm soil layers at the jointing stage, in the 40-80 cm and 140-160 cm soil layers at the silking stage, and in the 100-180 cm soil layers at physiological maturity than that under the treatments of N fertilization, respectively ( Figure 2E-H). Interestingly, with the extension of the growth period, the change in soil moisture among N treatments had an obvious trend to extend to the deep soil. The evapotranspiration varied with seasonal rainfall (Table 3). Evapotranspiration during the whole growth period of maize was greater by 52.6 mm in the 2020 growing season (wet) than that in the 2021 growing season (normal). In general, evapotranspiration was affected by different growth stages and N treatments (Table 3). On average, compared to CK, evapotranspiration with N fertilization treatments increased from sowing to jointing and from silking to maturity, but it was not significantly different from jointing to silking among different N managements. Under the same N fertilizer application rate, controlled-release N fertilization increased evapotranspiration from sowing to jointing in different degrees, but evapotranspiration from silking to maturity tended to decrease under controlled-release N fertilization. Compared to common urea application (0%CRU), The evapotranspiration varied with seasonal rainfall (Table 3). Evapotranspiration during the whole growth period of maize was greater by 52.6 mm in the 2020 growing season (wet) than that in the 2021 growing season (normal). In general, evapotranspiration was affected by different growth stages and N treatments (Table 3). On average, compared to CK, evapotranspiration with N fertilization treatments increased from sowing to jointing and from silking to maturity, but it was not significantly different from jointing to silking among different N managements. Under the same N fertilizer application rate, controlled-release N fertilization increased evapotranspiration from sowing to jointing in different degrees, but evapotranspiration from silking to maturity tended to decrease under controlled-release N fertilization. Compared to common urea application (0%CRU), on average, 30, 50, and 70% controlled-release N fertilization (i.e., 30%CRU, 50%CRU, and 70%CRU) significantly increased evapotranspiration from sowing to jointing by 22 mm under the same N fertilizer application rate. Compared to 0%CRU, 100% controlled-release N fertilization (100%CRU) significantly reduced evapotranspiration from silking to maturity by 19.5 mm (Table 3).

Effect of Controlled-Release N Fertilization on Soil Enzyme Activity and Soil Nitrogen Content
Soil nitrate nitrogen (NO 3 − -N), ammonium nitrogen (NH 4 + -N), and alkali-hydrolyzable nitrogen concentration and enzyme activities were profoundly affected by different nitrogen treatments (Figures 3 and 4, Table 4), but different nitrogen treatments did not affect soil total nitrogen content (Figures 3 and 4 and Table 4). At the silking stage, compared to the treatment without N (CK), N fertilization significantly increased alkalihydrolyzable nitrogen ( Figure 3B) and NO 3 − -N and NH 4 + -N concentration ( Figure 4A,B) in the 0-20 cm soil depth. However, under the same N fertilizer application rate, the greatest soil alkali-hydrolyzable nitrogen concentration and NO 3 − -N were found under common N fertilization (i.e., 0%CRU), followed by 30%CRU, and were lowest under 50%CRU, 70%CRU, and 100%CRU; 70%CRU significantly increased NH 4 + -N concentration in the 0-20 cm soil layer (Figures 3B and 4A).   Soil urease activity was greatest under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lowest among the treatment deprived of nitrogen (N) ( Table 4). The greatest soil nitrate reductase activity was found under 100%CRU, followed by 0%CRU, 70%CRU, 50%CRU, and 30%CRU, and was lowest under the treatment deprived of N. Under the same nitrogen fertilizer application rate, soil nitrite reductase activity was higher under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lower under 0%CRU. The 0%CRU and 30%CRU had the greatest soil sucrose activity, followed by 100%CRU, CK, 50%CRU, and 70%CRU. Nevertheless, there were no statistically significant distinctions observed in the soil catalase activity across the various nitrogen (N) treatments.   Soil urease activity was greatest under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lowest among the treatment deprived of nitrogen (N) ( Table 4). The greatest soil nitrate reductase activity was found under 100%CRU, followed by 0%CRU, 70%CRU, 50%CRU, and 30%CRU, and was lowest under the treatment deprived of N. Under the same nitrogen fertilizer application rate, soil nitrite reductase activity was higher under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lower under 0%CRU. The 0%CRU and 30%CRU had the greatest soil sucrose activity, followed by 100%CRU, CK, 50%CRU, and 70%CRU. Nevertheless, there were no statistically significant distinctions observed in the soil catalase activity across the various nitrogen (N) treatments.  Soil urease activity was greatest under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lowest among the treatment deprived of nitrogen (N) ( Table 4). The greatest soil nitrate reductase activity was found under 100%CRU, followed by 0%CRU, 70%CRU, 50%CRU, and 30%CRU, and was lowest under the treatment deprived of N. Under the same nitrogen fertilizer application rate, soil nitrite reductase activity was higher under 100%CRU, followed by 30%CRU, 50%CRU, and 70%CRU, and was lower under 0%CRU. The 0%CRU and 30%CRU had the greatest soil sucrose activity, followed by 100%CRU, CK, 50%CRU, and 70%CRU. Nevertheless, there were no statistically significant distinctions observed in the soil catalase activity across the various nitrogen (N) treatments.

Effect of Controlled-Release N Fertilization on Chlorophyll Content and Dry Matter Accumulation and Remobilization
The plant height of maize at the seedling stage was not significantly different among N treatments, but compared to the treatment without N, N fertilization significantly increased plant height at the jointing stage, silking stage, milk ripening stage, and maturity stage, respectively ( Figure 5). However, controlled-release N fertilization did not significantly affect plant height at all the growth stages compared to common N fertilization (i.e., 0%CRU). N fertilization increased the green color index (SPAD) compared to without N fertilization ( Figure 6). Compared to common N fertilization, controlled-release N fertilization slightly increased SPAD values, and the improvement in chlorophyll content showed an increasing trend with the extension of the growth period ( Figure 6A,B).

Effect of Controlled-Release N Fertilization on Chlorophyll Content and Dry Matter Accumulation and Remobilization
The plant height of maize at the seedling stage was not significantly different among N treatments, but compared to the treatment without N, N fertilization significantly increased plant height at the jointing stage, silking stage, milk ripening stage, and maturity stage, respectively ( Figure 5). However, controlled-release N fertilization did not significantly affect plant height at all the growth stages compared to common N fertilization (i.e., 0%CRU). N fertilization increased the green color index (SPAD) compared to without N fertilization ( Figure 6). Compared to common N fertilization, controlled-release N fertilization slightly increased SPAD values, and the improvement in chlorophyll content showed an increasing trend with the extension of the growth period ( Figure 6A,B). Biomass accumulation of maize varied with year, growing stage, and N treatments ( Table 5). Biomass accumulation in the 2020 season was greater at the seedling stage, jointing stage, and silking stage than in the 2021 season, but it was higher at the milk ripening stage and maturity stage in the 2021 season. Compared to the treatment without N, averaged across the five N application treatments, biomass accumulation under N fertilization was increased by 22% at the seedling stage, jointing stage, and silking stage, by 33% at the milk ripening stage, and by 37% at the maturity stage, respectively. Compared to U200, on average, controlled-release N fertilization increased biomass accumulation by 12% at the seedling stage, by 7% at the jointing stage, by 8% at the silking stage, by 11% at the milk ripening stage, and by 5% at the maturity stage, respectively; meanwhile, the proportion of biomass increase under 70%CRU was larger at all growth stages than that in the other treatments.  Biomass accumulation of maize varied with year, growing stage, and N treatments ( Table 5). Biomass accumulation in the 2020 season was greater at the seedling stage, jointing stage, and silking stage than in the 2021 season, but it was higher at the milk ripening stage and maturity stage in the 2021 season. Compared to the treatment without N, averaged across the five N application treatments, biomass accumulation under N fertilization was increased by 22% at the seedling stage, jointing stage, and silking stage, by 33% at the milk ripening stage, and by 37% at the maturity stage, respectively. Compared to U200, on average, controlled-release N fertilization increased biomass accumulation by 12% at the seedling stage, by 7% at the jointing stage, by 8% at the silking stage, by 11% at the milk ripening stage, and by 5% at the maturity stage, respectively; meanwhile, the proportion of biomass increase under 70%CRU was larger at all growth stages than that in the other treatments. Table 5. Effect of controlled-release urea fertilization on biomass accumulation of maize at the seedling stage, jointing stage, silking stage, milk ripening stage, and maturity stage during the 2020-2021 seasons.

Year
Treatment Biomass Accumulation (kg ha -1 )   Pre-anthesis dry matter remobilization (DMR), post-anthesis DM accumulation (PDM), the extent to which pre-anthesis dry matter is mobilized to the grain (CDMR), and the extent to which post-anthesis dry matter accumulates in the grain (CPDM) were affected by year and N managements (Table 6). DMR, PDM, CDMR, and CPDM in the 2020 growing season were lower than in 2021. On average, CDMR varied from 13.4% to 32.4%, and CPDM varied from 67.7% to 86.7%. The treatment without N fertilization significantly enhanced DMR and CDMR, but reduced PDM and CPDM when compared to N fertilization treatments (Table 5). Compared to 0%CRU, controlled-release N fertilization significantly increased DMR and CDMR. DMR and CDMR were the greatest in 50%CRU and were lowest in 0%CRU. Compared to 0%CRU, PDM in 70%CRU was significantly increased by 12%, but it was not significantly different in the other treatments. However, controlled-release N fertilization tended to decrease CPDM. Under 200 kg N ha −1 , the greatest CDMR was observed in 0%CRU, followed by 30%CRU and 70%CRU, and was lowest in 50%CRU and 100%CRU. Table 6. Effect of controlled-release urea fertilizer on pre-anthesis dry matter remobilization (DMR), the contribution of pre-anthesis dry matter remobilization to grain (CDMR), post-anthesis DM accumulation (PDM), and the contribution of post-anthesis dry matter accumulation to grain (CPDM).

Effect of Controlled-Release N Fertilization on Yield and Water Use Efficiency
Grain quantity per ear, 100-grain weight, and grain yield were lower in CK than in other treatments (Table 7). Under the same N application rate, compared to 0%CRU, grain number per ear in 70%CRU was significantly increased by 6.0% and 8.5% in the 2020 and 2021 seasons, respectively; 100-grain weight was increased by 9.2 and 10.8 (p > 0.05), respectively, in the 2020 and 2021 growing seasons. The quantity of grains per ear and the weight of 100 grains showed an average increase of 7.2% and 10.0%, respectively, under the 70%CRU treatment compared to the 0%CRU treatment. Grain yield was greater in 2020 (wet) than in 2021 (dry). Compared to 0%CRU, grain yield under 70%CRU was significantly increased by 30.1 and 11.9%, respectively, in the 2020 and 2021 growing seasons (p < 0.05). On average, grain yield was highest in 70%CRU, followed by 50%CRU, 100%CRU, 30%CRU, and 0%CRU, and was lowest in CK. The harvest index varied from 45.6% to 55.6%; on average, the HI was highest in 50%CRU and 70%CRU, followed by 30%CRU, and was lowest in CK, 100%CRU, and 0%CRU.
WUEg varied from 22.6 kg mm −1 ha −1 to 36.7 kg mm −1 ha −1 among different N treatments. The lowest WUEg was found in CK. Under the same N application rate, on average, WUEg tended to increase with the proportion of the controlled-release nitrogen fertilizer application. WUEg under 70%CRU was significantly increased by 17.2% compared to 0%CRU (Table 8). The trend of WUEg change with N management is consistent with the trend of WUEg change (Table 8).

Effect of Controlled-Release N Fertilization on Nitrogen Use Efficiency
The agronomic efficiency of nitrogen (AEN), apparent nitrogen recovery (ANR), and physiological use efficiency of nitrogen (PhEN) varied with year and N treatments (Table 9). On average, AE N was significantly increased by 23.5, 51.0, 90.8, and 25.5%, respectively, in 30%CRU, 50%CRU, 70%CRU, and 100%CRU when compared to 0%CRU (Table 9). ANR varied from 26.1% to 39.7%; on average, ANR was highest in 100%CRU, followed by 70%CRU and 50%CRU, and was lowest in 30%CRU and 0%CRU. The physiological use efficiency of nitrogen varied from 14.6 kg kg −1 to 83.8 kg kg −1 ; PhEN increased with an increase in the controlled-release urea application amount. Table 9. Effect of controlled-release urea fertilizer on the agronomic efficiency of nitrogen (AEN), apparent nitrogen recovery (ANR), and physiological use efficiency of nitrogen (PhEN) during the 2020-2021 seasons.

Effect of Controlled-Release N Fertilization on Soil Enzyme Activities, Biomass Accumulation, and Grain Yield
Soil enzyme activities were greatly increased under N fertilization, mainly because fertilizer application regulated the activities of enzymes involved in nutrient turnover's utilization of substrate [39,40]. At the silking stage, controlled-release nitrogen fertilization increased the associated enzymes involved in soil nitrogen utilization, such as soil urease, nitrate reductase, and nitrite reductase, to variable degrees, consistent with a previous study [41]. It is very important to improve the soil nitrogen supply and promote the nitrogen uptake of maize. However, under controlled-release urea fertilization, compared to conventional urea treatment, there was no increase in the content of soil nitrate nitrogen content and alkali-hydrolyzable nitrogen. This was likely because high biomass accumulation enhanced N uptake, thus reducing soil nitrate nitrogen and alkali-hydrolyzable nitrogen concentrations in controlled-release urea fertilizer treatments.
In dryland regions, the productivity of crops is significantly influenced by both soil water conditions and fertilization practices [4,19]. Nitrogen fertilization increased grain yield, mainly because of an increase in biomass accumulation, especially post-anthesis dry matter buildup, thus improving the grain no. per ear and grain weight, which is in line with earlier research [4,19,31]. Our two-year study showed that controlled-release N fertilization increased grain yield. Similar findings were discovered in wheat and maize cropping seasons [2,10,23]. Grain yield was increased with the proportion of the controlled-release nitrogen fertilizer application increase, mainly because of a greater biomass accumulation during the growth period and pre-anthesis dry matter remobilization. High biomass accumulation and pre-anthesis dry matter remobilization increased yield components, such as the grain no. per ear and 100-grain weight, as it has been shown in previous stud-ies [4,19,38,42]. In this study, grain yield first increases and then decreases with an increase in the proportion of controlled-release nitrogen fertilization; this was likely because the increases in biomass accumulation and pre-anthesis dry matter remobilization were not accompanied by an increase in the proportion of controlled-release nitrogen fertilizer application, When the proportion of controlled-release nitrogen fertilizer application exceeds 70%, the accumulation of biomass showed a downward trend, which is in agreement with results from a previous study [41].

Effect of Controlled-Release N Fertilization on Water and Nitrogen Use Efficiency
High biomass accumulation under controlled-release urea, especially biomass accumulation during the vegetative growth stage, greatly enhanced WUEb, thereby improving WUEg. Similar findings were discovered in the wheat growing season [4,7]. This is likely because the greater biomass accumulation before anthesis increases canopy cover, reduces canopy temperature, and enhances crop transpiration efficiency [38,43,44]. This result is supported by our evidence that although controlled-release urea application significantly increased biomass accumulation, it did not increase evapotranspiration. The trend of changes in the agronomic efficiency of nitrogen is similar to WUEg. Controlled-release nitrogen fertilization increases the nitrogen use efficiencies of maize, which is in agreement with results from previous studies [16,23,27]. As the amount of controlled-release urea applied increases, the agronomic efficiency of nitrogen initially improves and then declines. Maize has shown similar responses in previous studies [10,25,27]. The high agronomic efficiency of nitrogen was associated with an increase in biomass accumulation, grain no. per ear, 100-grain weight, and harvest index. High N uptake under controlled-release urea enhanced biomass accumulation, thus increasing apparent nitrogen recovery. Increased biomass accumulation and grain yield improved the physiological use efficiency of nitrogen, ultimately leading to an increase in the agronomic efficiency of nitrogen. These findings suggest that there is an optimal proportion of controlled-release urea application that can enhance the efficiency of nitrogen in maize farming in the semiarid region of the Loess Plateau. Taking all the factors into consideration, we conclude that a combination of 70% controlled-release urea and 30% common urea should be considered as the suitable application ratio in the semiarid region of Loess Plateau within western China.

Conclusions
The present findings illustrate the significant prospects for optimizing the N fertilizer management for spring maize on the semiarid Loess Plateau. This study found that 70%CRU combined with 30% common urea had great potential for improving the resource utilization efficiency of maize under plastic film mulching on the semiarid Loess Plateau. Nevertheless, the expensive cost of controlled-release urea also has limitations on the popularization of this fertilizer. In the future, we should pay more attention to the low-cost production of controlled-release urea for improving its promotion, and multi-locational studies are also suggested to clarify how climatic variations in different semiarid regions would influence the potential effects of N fertilization on maize yield and water and nitrogen utilization efficiency.