Deep Fertilization Is More Beneficial than Enhanced Efficiency Fertilizer on Crop Productivity and Environmental Cost: Evidence from a Global Meta-Analysis

Abstract

It is unclear whether enhanced efficiency fertilizer (EEF) or deep fertilization strategies (DF) can simultaneously improve crop productivity and reduce gaseous nitrogen losses. The DF strategy’s investment cost is lower than that of EEF’s, with more potential for large-scale promotion. However, there is still a need for a comprehensive comparison and evaluation of DF and EEF’s effects on crop productivity and gaseous nitrogen losses. Here, we examine the effects of DF and EEF on crop yield, nitrogen use efficiency (NUE), and nitrous oxide (N₂O) and ammonia (NH₃) emissions by a meta-analysis of published studies. We collected peer-reviewed articles on EEF and DF published in recent decades and conducted a global meta-analysis, and explored their responses to different climatic, field management practices, and environmental factors. The results showed that compared with urea application on the surface, EEF and DF significantly increased yields by 7.52% and 13.88% and NUE by 25.84% and 36.27% and reduced N₂O emissions by 37.98% and 34.18% and NH₃ emissions by 42.37% and 69.68%, respectively. The DF strategy is superior to that of the EEF. Due to differences in climatic factors, soil properties, and management practices, the effects of DF and EEF in improving crop productivity and gaseous nitrogen loss vary. However, in most cases, DF is more beneficial than EEF. Compared with EEF, DF significantly increased the yield by 84.63% and reduced NH₃ volatilization by 64.47%, yield-scaled N₂O emission by 13.32%, and yield-scaled NH₃ emission by 60.23%. Therefore, we emphasize that DF can achieve higher yields, nitrogen fertilizer utilization efficiency, lower emissions of gaseous nitrogen, and lower yield-scaled N₂O and NH₃ emissions than EEF, which is beneficial for the sustainable development of global agricultural ecosystems. The research results provide valuable information on crop productivity and environmental costs under an effective fertilizer type and fertilization strategy management.

1. Introduction

The large gap between the demand for food and the production of food is increasing, and the reduction in food production caused by climate change is intensifying [1,2]. Agricultural activities have contributed significantly to global food security and climate change [3], including the extensive application of chemical fertilizers and intensive livestock farming on grasslands [4]. Ammonium (NH₃) and N₂O have polluted the environment, posing a threat to global public health [5,6,7]. Excessive fertilizer input and improper fertilization methods result in nitrogen loss in the form of runoff, leaching, NH₃ volatilization, and N₂O emissions, significantly reducing nitrogen fertilizer utilization efficiency and potential economic benefits in farmland [8,9]. Therefore, optimizing fertilizer management strategies to produce more food per unit land area, reduce NH₃ and N₂O emissions, and promote sustainable agriculture development has attracted significant global interest.

The 4R principle (right fertilization type, right fertilization place, right fertilization rate, and right fertilization time) was proposed to achieve a win–win goal of increasing crop productivity and reducing nitrogen loss, and it has been widely recognized [3,10]. Reducing fertilizer input may lead to lower nitrogen loss but also risks decreasing crop yields [11,12]. Splitting fertilization during the critical growth period of crops improves nitrogen utilization efficiency [13,14] but increases N₂O and NH₃ emissions, labor demand, and the cost of inputs [10,15]. Therefore, farmers still tend to adopt one-time fertilization technology in agricultural production. Using fertilizer types and fertilization strategies that can increase economic benefits and ecological news may change farmers’ choices.

Recent studies have shown that the application of enhanced efficiency fertilizer (EEF)and deep fertilization (DF) strategies may be more suitable for agricultural ecological environments [16,17,18]. EEF synchronizes the nitrogen supply and crop uptake by regulating the nitrogen conversion process [19]. Current types of EEFs include controlled-release urea (CRF) and the addition of nitrification inhibitors (NIs), and urease inhibitors (UIs), as well as a combination of these two inhibitors in nitrogen fertilizer (UI + NI) [20,21,22]. Controlled-release urea mainly achieves a slow release of nitrogen through coating technology to match the demand pattern for nitrogen during crop growth [23]. NIs mainly reduce N₂O emissions by inhibiting the conversion process of NH₄⁺ to NO₃⁻ [24], while UIs mainly delay the hydrolysis of urea by inhibiting urease activity to reduce NH₃ volatilization [25]. The combined application of NIs and UIs can simultaneously reduce NH₃ and N₂O emissions. However, due to the existence of ”pollution exchange”, the cost of reducing the emissions of one pollutant is the increase in the emissions of another pollutant [26]. In addition, the impact of controlled-release urea on NH₃ and N₂O emissions varies [27,28]. An increasing number of studies have shown that the research results of EEF vary greatly in different regions, and its effectiveness is greatly influenced by soil characteristics, climate conditions, and management practices [29]. These contradictory results indicated that the impact of EEF on farmland ecology is highly complex. Therefore, there is an urgent need to obtain more general conclusions through a comprehensive analysis of global research data to guide production practice.

DF is considered an important fertilization strategy for improving NUE and yield [30,31]. However, in the past, due to the low level of mechanization, manual deep fertilization has been used, making it difficult to promote DF technology. In recent years, with the development of global agricultural mechanization, mechanical deep fertilization has been widely used in agricultural production [32,33]. Some studies have found that deep fertilization can significantly improve crop yield and nitrogen utilization efficiency [34] and reduce NH₃ volatilization [32]. However, the impact of deep fertilization on N₂O emissions is still unclear. Wu et al. [35] found that deep fertilization can significantly reduce N₂O emissions, but Maaz et al. [36] found that deep fertilization increases the N₂O emissions from farmland. Therefore, the impact of deep fertilization on farmland ecosystems may be influenced by climate factors, soil properties, and management measures. At present, there is a lack of a comprehensive evaluation of the effects of deep fertilization on crop yield, nitrogen fertilizer utilization efficiency, and N₂O and NH₃ emissions from farmland, and key regulatory factors still need to be further elucidated.

Although EEF and DF are beneficial for improving the sustainability of farmland ecosystems, the high price of EEF restricts their application to some extent [37]. There has been limited research comparing the effects of DF and EEF on crop yield, NUE, and N₂O and NH₃ emissions. A key question remains in that we do not know the different effects of DF and EEF on crop productivity and gaseous nitrogen losses, and, currently, there is limited research on this topic. For this reason, we collected 4463 pairs of experimental data from 371 published literature studies worldwide, using traditional fertilization as the control, that is, with a traditional fertilization depth and without the addition of inhibitors, and deep fertilization or the application of inhibitors as the treatment. Our aims were the following: (1) to evaluate the impact of EEF and DF on crop productivity and environmental sustainability; (2) to identify the driving factors that related to crop productivity and environmental sustainability under the use of EEF and DF strategies and to quantify the impact of these different factors; (3) to analyze and compare the advantages, disadvantages, and applicability of EEF and DF in the sustainable development of global farmland ecosystems and to determine the best management practices for the sustainable development of farmland ecosystems; (4) to deepen our understanding of EEF and DF strategies and provide a reference for global food security and climate change research.

2. Materials and Methods

2.1. Literature Search and Study Selection

In this meta-analysis, we selected two management practices, namely EEF and DF, and retrieved data from articles published before March 2023 from the Google Scholar (http://scholar.google.com/ (accessed on 10 March 2023)), Web of Science (http://apps.webofknowledge.com/ (accessed on 10 March 2023)), and China National Knowledge (https://www.cnki.net/ (accessed on 10 March 2023)) databases to explore the impact of enhanced efficiency fertilizer and deep fertilization on crop productivity and the environment. The keyword panels we selected were (i) “enhanced efficiency fertilizer”, “enhanced efficiency nitrogen”, “slow-release nitrogen”, “controlled-release nitrogen”, “nitrification inhibitors”, “urease inhibitors”, “deep fertilization”, “deep placement fertilizer”, “deep placement nitrogen”, “fertilization depth”, “urea deep placement” and (ii) “yield”, “nitrogen use efficiency”, “NUE”, “nitrous oxide”, “N₂O”, “ammonia volatilization”, “NH₃”, “reactive nitrogen”, “gaseous nitrogen”, “yield scaled- N₂O emission”, “yield scaled-NH₃ emission”.

An article had to meet the following conditions before being included in the database: (1) All the research must have been conducted on-site; (2) The experiment must have used ordinary urea or shallow fertilization as the control group and synergistic fertilizer or deep fertilization as the treatment group, and conducted paired comparisons; (3) The experimental data must include the average, standard deviation or standard error, and number of replicates of our target parameters (yield, NUE, N₂O, or NH₃). The distribution of studies using the enhanced efficiency fertilizer strategy and deep fertilization strategy included in the global meta-analysis are shown in Figure 1.

2.2. Data Extraction

We extracted the mean, number of replicates, standard error, or standard deviation of the target parameters for each study’s control and treatment groups. If the article did not show the standard error or standard deviation, we assumed that the standard deviation was 10% of the average of the data. The data presented in the literature were directly obtained in tabular form. If presented in graphical form, we used GetData Graph Digitizer 2.26 (https://getdata-graph-digitizer.com/ (accessed on 16 September 2022)) to extract the data.

To investigate the key factors related to crop yield, NUE, and N₂O and NH₃ emissions, and to investigate the effects of EEF and DF, we recorded a series of environmental and experimental variables: location (longitude and latitude), climate (annual average temperature and precipitation), soil conditions (soil pH, total nitrogen, organic carbon, soil texture), and management practices (fertilizer type, fertilization depth, nitrogen application rate, crop type, etc.). For some missing information, we extracted MAT and MAP from WorldClim2.1 (https://www.worldclim.org/ (accessed on 7 January 2024)), and soil content from SoilGrid2.0 (https://soilgrids.org/ (accessed on 7 January 2024)). In addition, to further analyze the impact of different factors on the research parameters, we divided the climate, environmental variables, and management practice databases into sub databases (

Table 1. Specific classification and number of data pairs for each subgroup. MAT and MAP represent the mean annual temperature and mean annual precipitation; TN and SOC represent the total soil nitrogen and soil organic carbon, respectively; UI, NI, and CRU represent urease inhibitor, nitrification inhibitor, and controlled-release nitrogen, respectively.

Categorical Variables		Groups
Climate factor	MAT (°C)	<5	5–10	10–15	15–20	>20
	MAP (mm)	<400	400–800	800–1200	>1200
Soil factor	TN (g·kg−1)	<0.7	0.7–1.4	1.4–2.1	>2.1
	pH	<6	6–7	7–8	>8
	SOC (g·kg−1)	<10	10–20	20–30	>30
	Soil texture	Fine	Medium	Coarse
Management practice factor	Nitrogen rate (kg·ha−1)	<150	150–225	225–300	>300
	Fertilizer type	UI	NI	UI + NI	CRU
	Fertilization depth (cm)	<5	5–15	15–25	>25

).

We collected the yield-scaled N₂O emission and yield-scaled NH₃ emission data from the article, or used Formulas (1) and (2) for the calculations.

$$\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d} \mathrm{s}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{d}-{\mathrm{N}}_2\mathrm{O} \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}={\displaystyle \frac{{\mathrm{N}}_2\mathrm{O}}{\mathrm{C}\mathrm{r}\mathrm{o}\mathrm{p} \mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}}}$$

(1)

$$\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d} \mathrm{s}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{e}\mathrm{d}-{\mathrm{N}\mathrm{H}}_3 \mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}={\displaystyle \frac{{\mathrm{N}\mathrm{H}}_3}{\mathrm{C}\mathrm{r}\mathrm{o}\mathrm{p} \mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}}}$$

(2)

2.3. Data Analysis

The impact of EEF or DF on the yield, NUE, and N₂O, and NH₃ emissions was evaluated by calculating the natural logarithm of the response ratio. All the data conformed to a normal distribution (Figure S1). A linear mixed-effect model with the “rma.mv” function in the R4.1.1 package “metafor” was used to calculate the overall effect value. The mixed-effects model considers “study” and “observation” as random factors to ensure the independence of each observation. The percentage change in the treatment group [(e^lnR − 1) × 100%] represents the change result, where a positive result indicates an increase in the target parameter and a negative change indicates a decrease. The average effect and the 95% confidence interval were calculated using the “metafor” function of the R4.1.1 package. If the 95% confidence interval did not include the value 0, it was considered that the effect of EEF and DF on the target parameters was significant. In analyzing different parameters (yield, NUE, N₂O, yield-scaled N₂O emission, NH₃, yield-scaled NH₃ emission), we divided DF and EEF into two groups. Significant heterogeneity between the groups meant a substantial difference in the average effect size between the DF and EEF groups. Please refer to Tables S1–S3 in the Supplementary Materials for the analysis results.

We used the mixed-effect regression model in the R4.1.1 package “glmulti” to select and analyze the most important predictive factors used to determine the impact of EEF or DF on the target parameters. The model selection was based on maximum-likelihood estimation. We considered the sum of Akaike weights predicted by the model as the importance of the predictive factors and used a critical value of 0.8 to distinguish between important and non-important predictive factors. Also, we included all available predictive factors (MAP, MAT, SOC, TN, pH, soil texture, latitude, crop species, fertilization type, fertilization depth, nitrogen rate) in the model’s selectable variables and used regression analysis to study the relationship between the ln R of the target variable and the variables.

3. Results

3.1. Overall Effect

EEF and DF significantly increased crop productivity and decreased gaseous nitrogen losses. The yield- and NUE-increase effects and gaseous nitrogen loss reduction effects under DF were more significant than those under EEF (Figure 2). Compared with the controlled treatments, EEF significantly increased the yield and NUE by 7.52% (95% CI: 6.27–8.80%) and 25.83% (95 CI: 20.47–31.44%), while DF significantly increased the yield and NUE by 13.88% (95 CI: 10.83–17.03%) and 36.27% (95 CI: 23.45–50.44%). In addition, DF increased the yield and NUE by 84.63% and 40.41% compared with EEF.

EEF significantly decreased N₂O emissions and yield-scaled N₂O emissions by 37.98% (95% CI: 32.63–42.91%) and 42.37% (95% CI: 40.25–44.49%), and DF significantly decreased them by 34.18% (95% CI: 16.32–48.23%) and 69.68% (95% CI: 29.78–52.71%). Compared with EEF, DF increased N₂O emissions by 10.00% but decreased NH₃ emissions by 64.47%.

Also, we discovered a significant decrease in yield-scaled N₂O and yield-scaled NH₃ emissions of 39.67% (95% CI: 25.85–59.14%) and 45.25% (95% CI: 53.34–83.79%) under EEF treatment. DF significantly decreased the yield-scaled N₂O and yield-scaled NH₃ emissions by 44.96% (95% CI: 33.49–45.29%) and 72.50% (95% CI: 30.52–56.85%). DF decreased the yield-scaled N₂O and yield-scaled NH₃ emissions by 64.47% and 60.23%, respectively, compared with EEF.

3.2. Effects of Climate Characteristics

MAT and MAP were significantly related to the yield- and NUE-increase effect, as well as to the gaseous nitrogen loss reduction effect (Figure 3). In terms of yield and NUE, when MAT > 10 °C or MAP > 400 mm, both EEF and DF significantly increased the yield and NUE. When MAT > 20 °C or MAP > 1200 mm, EEF and DF achieved the highest yield and NUE effects. The increase effects on the yield and NUE by EEF were 10.54% and 23.78% and those by DF were 25.53% and 45.92%, respectively, when MAT > 20 °C, and 10.54% and 32.15% under EEF and 17.13% and 57.90% under DF when MAP > 1200 mm. In addition, DF obtained the highest N₂O emission-reduction effects of 56.96% when MAT was within 15–20 °C and 47.24% when MAP > 1200 mm. However, no significant differences were observed in the N₂O emission and yield-scaled N₂O emission under EEF and DF. DF significantly decreased the NH₃ emission and yield-scaled NH₃ emission by 81.49% and 86.32% when MAP was within 800–1200 mm, which was significantly lower than that seen with EEF.

3.3. Effects of Soil Properties

EEF and DF significantly increased the yield regardless of different soil properties, except for soil pH > 8 under DF conditions (Figure 4). The yield-increasing effects of EEF and DF rise with SOC. The effects of EEF remain relatively stable with pH, while those of DF decline as the pH increases. DF obtained the highest yield-increase effects when the soil TN > 2.1 g kg⁻¹ (19.63%), pH < 6 (17.76%), and 20 g kg⁻¹ < SOC < 30 g kg⁻¹ (19.36%) compared with EEF (p < 0.05).

Both EEF and DF significantly increased the NUE, but no significant difference was observed in the NUE-increase effects under DF and EEF regardless of soil pH, SOC, and soil texture. DF significantly increased the NUE (78.82%, 95% CI: 53.08–108.86%) when 1.4 < TN < 2.1 g kg⁻¹ compared with EEF (p < 0.05).

DF and EEF significantly decreased the N₂O emission and yield-scaled N₂O emission, but there was no significant difference between EEF and DF. DF obtained the highest N₂O emission (45.40%, 95%CI: 14.18–65.26%) and yield-scaled N₂O emission (66.07%, 95%CI: 36.93–81.75%) reduction effects when the soil texture was fine.

Regardless of soil properties, DF and EEF both reduced the NH₃ emission and yield-scaled N₂O emission. However, regardless of soil pH, there were no significant differences between EEF and DF on NH₃ emission and SOC on yield-scaled NH₃ emission. The NH₃ emission (70.62%, 95% CI:55.01–80.81%) and yield-scaled NH₃ emission (76.54%, 95% CI: 57.48–87.05%) reduction effects under DF were significantly higher than under EEF when 1.4 g kg⁻¹ < TN < 2.1 g kg⁻¹. In addition, the NH₃ emission-reduction effects under DF were significantly higher than under EEF when 10 g kg⁻¹ < SOC <20 g kg⁻¹ and the soil texture was coarse.

3.4. Effects of Field Management Practices

DF and EEF significantly increased the yield and NUE regardless of the nitrogen application rate (Figure 5). When the nitrogen application rate was between 150–225 kg ha⁻¹ and above 225 kg ha⁻¹, the yield-increase effect of DF was significantly higher than that of EEF, with an increase of 97.66% and 109.45%, respectively. The yield-increase effects (14.82%, 95%CI: 10.77–19.02%) and NUE-increase effects (39.39%, 95%CI: 22.37–58.77%) were the highest under DF when the nitrogen rate was within 150–225 kg ha⁻¹. EEF obtained the highest NUE-increase effects when the EEF species was an NI.

Regarding N₂O and yield-scaled N₂O emissions, there was no significant difference between DF and EEF, but DF obtained higher yield-scaled N₂O emission-reduction effects than EEF (Figure 2, p > 0.05). When the fertilization depth reached more than 25 cm, the effect on N₂O emission reduction was the best. The EEF combination of UI + NI obtained the highest N₂O emission-reduction effects, but NIs obtained the highest yield-scaled N₂O emission-reduction effect.

DF obtained higher NH₃ emission and yield-scaled NH₃ emission effects than EEF, regardless of field management practices. When the nitrogen application rate was within 225–300 kg ha⁻¹ and above 300 kg ha⁻¹, DF had the highest yield-scaled NH₃ emissions of 85.13% and 92.38%, significantly higher than those under EEF. DF obtained the highest NH₃ emission and yield-scaled NH₃ emission-reduction effects when the fertilization depth was over 25 cm. There were no significant reduction effects on NH₃ emission and yield-scaled NH₃ emission when the EEF species was an NI.

3.5. Correlation Between Climate Characteristics, Soil Properties, Field Management Practices, Crop Productivity, and Gaseous Nitrogen Losses

From the perspective of climate characteristics, the increased effects on the yield and NUE and the reduction effects on N₂O emission, yield-scaled N₂O emission under EEF increased with an increase in MAT and MAP and decreased with an increase in latitude (Figure 6). However, the reduction effects on NH₃ emission and yield-scaled NH₃ emission decreased with an increase in MAP and MAT under EEF.

In terms of soil properties (Figure 7), under EEF conditions, the increase effect on the yield increased with an increase in SOC. However, the reduction effect on N₂O emissions and N₂O emissions per unit yield shows a trend of first decreasing and then increasing with the increase in SOC. The ln R of NH₃ and yield-scaled NH₃ emission under EEF and DF increased with soil pH. However, under EEF conditions, the ln R value of N₂O emissions shows a trend of increasing and then decreasing with the increase in soil pH. The ln R of N₂O, NH₃ emission, yield-scaled N₂O, and NH₃ emission decreased with the increase in soil TN.

In terms of field management practices (Figure 8), the increasing effect on the yield under the DF condition first increased and then decreased with the increase in the nitrogen application rate. However, the reduction effect on N₂O increased, and the reduction effect on yield-scaled NH₃ emission decreased first and then increased. The increasing effect on the NUE under EEF decreased with the increase in the nitrogen application rate. The reduction effect on NH₃ emission increased with the increase in fertilization depth.

3.6. Important Predictors

Model selection analysis showed that the most important factors related to the yield, NUE, and N₂O and NH₃ emission under EEF and DF were different (Figure 9). Overall, the fertilizer type was the most important factor related to the N₂O, NH₃, and yield-scaled N₂O and NH₃ emission under EEF. The nitrogen rate was the most important factor related to the N₂O emission and yield-scaled N₂O emission under DF, and the fertilization depth was the most important factor related to the NH₃ emission and NUE under DF. The crop species and soil texture were the most important factors related to the yield under EEF, while under DF conditions, the main factors related to yield were latitude, soil TN, and soil pH.

4. Discussion

The results of this study, collected from across six continents worldwide, indicate that EEF and DF have high universal applicability in improving global crop productivity and reducing gaseous nitrogen loss. Our results again confirm that the 4R principle of EEF and DF can achieve higher economic benefits with lower environmental costs. DF significantly increased the crop yield and reduced NH₃ volatilization compared with EEF. In addition, our research results indicated no significant difference in the role of EEF and DF in most climate, soil properties, and management practice situations. However, when MAT > 20 °C, TN > 2.1 g kg⁻¹, pH < 6, and the nitrogen rate < 150 kg ha⁻¹, the yield increase and emission-reduction effect of DF were significantly higher than those of EEF, indicating that DF is more beneficial than EEF globally.

4.1. General Effects of DF and EEF on Crop Productivity

Our research results indicated that EEF and DF can increase crop yield and NUE. However, DF has a stronger ability to improve crop productivity than EEF. In addition, EEF and DF can regulate the inorganic nitrogen content in the soil, which is important for absorption and utilization [38,39,40]. Rice mainly absorbs NH₄⁺-N. Due to rice seedlings’ low nitrogen demand, a large amount of NH₄⁺-N is lost after applying urea. UIs can inhibit the hydrolysis rate of urea for plant absorption and utilization [41,42]. NIs increase the NH₄⁺-N content in the soil by inhibiting the conversion of NH₄⁺-N to NO₃⁻-N, which is used for crop absorption and utilization [43,44]. CRF uses physical coating materials to delay the release of nitrogen, matching the nitrogen demand during crop growth and improving crop yield and fertilizer utilization efficiency [45]. As the latitude increases, the ln R of EEF gradually decreases. The main reason for this phenomenon is that as the latitude increases, the global temperature and precipitation decrease [45], making it difficult for EEFs to function. This is consistent with our finding that ln R was significantly positively correlated with MAT and MAP.

On the other hand, good soil nutrient content helps promote root development, thereby increasing crop yield [46,47]. Interestingly, we found that DF has a stronger ability to improve yield and NUE than EEF. The main reason is that DF has a greater impact on soil root development than EEF. Although both EEF and DF can alter the inorganic nitrogen content in the soil to promote root development [48], EEF mainly changes the form of the nitrogen fertilizer. At the same time, DF includes the deep application of nitrogen fertilizer, phosphorus fertilizer, and the combination of nitrogen and phosphorus fertilizer [48,49,50,51]. DF may have a greater impact on soil microbial activity in the root zone than EEF [52], which, to some extent, explains the stronger yield-increasing effect of DF compared with EEF. In addition, EEF generally adopts conventional fertilization methods, which involve spraying fertilizer on the surface and rotating it into the soil at shallow depths. At the same time, crop roots are mainly distributed in the soil at 0–60 cm depths. Surface nitrogen enrichment reduces root penetration [49]. Research has shown that moderate nitrogen deficiency can stimulate roots to penetrate deeper into the soil to explore nitrogen [53,54]. Deep fertilization can reduce the nitrogen content of the surface soil, increasing the nitrogen content of deep soil [55,56], thereby promoting root growth [51]. Our previous study also showed that DF can promote root development, delay leaf senescence, and improve nitrogen fertilizer utilization efficiency and crop yield [31].

Recent studies have also found that EEF and DF with nitrogen, phosphorus, or nitrogen and phosphorus fertilizers can delay the senescence of post-flowering leaves and promote grain formation [57,58]. However, research has found that nitrogen and phosphorus play vital roles in regulating leaf senescence [59,60]. Therefore, DF delays leaf senescence by applying nitrogen and phosphorus, significantly increasing crop yield and nitrogen uptake and ultimately increasing NUE [31].

4.2. General Effects of DF and EEF on the Reduction of Gaseous Nitrogen

Our research results indicated that DF and EEF can significantly reduce NH₃ and N₂O emissions. This is because DF and EEF can maintain more nutrients in the root zone and reduce nitrogen environmental losses [61]. As for UIs, these mainly reduce the rate of urea decomposition by inhibiting urease activity in the soil [62], thereby reducing NH₃ emissions. This study found that UIs’ NH₃ emission-reduction ability was stronger than that of DF. This is because the urease inhibitor (NBPT) is a urea analogue that is converted into N-(n-butyl) phosphate tripamide in aerobic soil and binds at the urease site, thereby slowing down urea hydrolysis and reducing NH₃ volatilization [22], which is a form of biochemical inhibition. However, DF reduces NH₃ emissions by reducing the nitrogen content of near-surface soil, which involves physical suppression. In terms of NIs, these directly target soil ammonia monooxygenase, slowing down the conversion of NH₃ to nitrite by microorganisms, which, in turn, is converted into NO₃⁻, reducing N₂O emissions [63]. Specifically, urea dissolves in the soil and quickly distributes in the soil, while positively charged NIs can adsorb on floating point soil colloids, effectively suppressing N₂O emissions [64]. The regulatory mechanism of NIs leads to the loss of nitrogen in the form of NH₃ volatilization [65]. In this study, it was discovered that NIs increase NH₃ emissions. Although NIs have a slightly stronger ability to reduce N₂O emissions than DF, DF has a significantly higher ability to reduce NH₃ volatilization than NIs. The combination of NIs and UIs can, to some extent, compensate for the NI’s disadvantage in increasing NH₃ emissions, significantly reducing NH₃ and N₂O emissions, but DF’s ability to reduce N₂O is still considerably stronger than that of the UI + NI combination. In terms of CRU, this controls nitrogen release through physical barriers or polymer soil layers, thereby synchronizing plant nitrogen demand and fertilizer nitrogen supply. Although CRU has a slightly stronger N₂O emission-reduction ability than DF, its NH₃ emission-reduction ability is significantly weaker than that of DF. The main reason for this phenomenon is that the nitrogen release rate is influenced by external factors such as soil temperature, moderation, and pH [66,67]. In addition, the coating material of CRU is expensive, and the release kinetics of CRU are difficult to understand in specific soil environments and conditions [68], and it cannot directly respond to the nutritional needs of plants. Therefore, the emission-reduction effect of DF is stronger than that of EEF.

The following points will help to increase our understanding of this phenomenon. Firstly, the emission reduction by EEF mainly relies on chemical reactions to prevent urea hydrolysis or NH₃ conversion. External factors often influence this chemical process. Some studies have reported that when the soil temperature increases, the inhibitory ability of NIs will sharply decrease, making it difficult to exert their inhibitory effects [64]. Nitrogen is the primary substrate for N₂O production in NH₃ [69]. After deep fertilization, the nitrogen content in the surface soil decreases, which, ultimately, reduces N₂O and NH₃ emissions [51]. Secondly, although EEF can prevent urea hydrolysis or NH₃ conversion, the traditional fertilization depth is relatively shallow, and N₂O and NH₃ will quickly migrate and discharge upward after being produced in the soil, ultimately leading to the loss of gaseous nitrogen [70]. Previous studies report that the 7–15 cm soil layer is the central location for N₂O production [71,72]. After deep application of fertilizers, the upward migration path length of N₂O and NH₃ is increased through physical means. During the migration process, N₂O is reduced to N₂ by various factors such as soil temperature, moisture, and O₂ content, which significantly lowers N₂O emissions [55,73]. Therefore, DF has fewer limiting factors in reducing gaseous nitrogen emissions, and its application advantages are stronger than those of EEF.

The yield-scaled N₂O emission and yield-scaled NH₃ emission are considered the primary economic yield benefits, and the environmental costs due to NH₃ and N₂O emissions as the leading yield indicators [74,75], which can be used to comprehensively evaluate the applicability of our management practices [76]. Our research results indicated that DF significantly reduces the yield-scaled N₂O emission (64.47%) and yield-scaled NH₃ emission (60.23%) compared with EEF, indicating that the environmental cost of producing the exact grain under DF is significantly lower than that under EEF. Therefore, DF’s potential to reduce the social cost of nitrogen pollution is higher than that of EEF, which has excellent potential for application globally.

4.3. Conditions in Which the DF Strategy More Beneficial than the EEF Strategy

Climate factors (MAP and MAT) regulate the effects of EEF and DF on crop productivity and gaseous nitrogen emissions. The impact of EEF on yield and NUE is not related to MAP and MAT, but DF is significantly associated with them. This phenomenon may be due to their different mechanisms in increasing production and NUE.EEF mainly regulates nitrogen release rate through biochemical means to retain more nitrogen in the soil to match crop growth patterns [77,78], especially CRU [79], so it is less related to MAT and MAP. When MAP and MAT increase or decrease, changes in the soil moisture, nutrients, and microbial activity can affect the efficacy of EEF [80], thereby regulating the yield and NUE. However, the yield-increasing effect of EEF is relatively stable. The increase in yield under the DF condition is due to the better absorption of nutrients by the root system, which, ultimately, leads to an increase in yield and NUE [81,82,83,84]. Various factors influence deep fertilization, such as the degree of matching between the fertilization depth and root distribution [84], the loss of nitrogen in the form of NH₃ and N₂O volatilization when fertilization is shallow [38], and the possibility of nitrogen leaching during the rainy season. As reported by Wang et al. [56], the optimal fertilization depth for wheat is significantly affected by the amount of irrigation during its growth period. This study found that when MAT < 10 °C and MAP < 400 mm, the improvement effect of DF on the yield and NUE was not as good as that of EEF. EEF mainly reduces NH₃ or N₂O emissions by regulating soil nitrification and denitrification processes [85], which are significantly influenced by water and temperature. We found that EEF is more related to MAT and MAP when reducing NH₃ emissions than N₂O emissions, which may be influenced by the inhibitor type [86]. In most cases, the emission-reduction effect of DF on NH₃ emission is significantly higher than that of EEF, as NIs can reduce N₂O emission but increase NH₃ emission. Although DF is affected by MAP and MAT and shows instability, its yield-increasing and emission-reducing effects are more potent than EEF’s.

The physical and chemical properties of soil are significantly related to the yield-increase and emission-reduction effects of EEF and DF due to their influence on the transformation process of nitrogen in the soil [87,88]. Our research results indicated that under most soil properties, such as 0.7 g kg⁻¹ < TN < 1.4 g kg⁻¹, pH < 6, and 20 g kg⁻¹ < SOC < 30 g kg⁻¹, the yield-increase effect of DF was significantly higher than that of EEF. In addition, DF has a stronger ability to reduce N₂O and NH₃ emissions than EEF, indicating that DF will have significant space for application in the future and can help alleviate global climate change.

Field management practices significantly affect crop yield and gaseous nitrogen emissions [89,90]. Overall, DF caused a greater increase in yield and NUE than EEF. Still, EEF’s yield-increase effect was relatively stable, and management practices significantly influenced DF. In this study, we discovered that the effects of DF and EEF on crop yield are unrelated to the three primary crop types (wheat, maize, and rice) and can all improve crop yield. However, the impact of EEF on potato and cotton yields was relatively small, mainly because wheat, maize, and rice are the main food crops, and a large amount of sample data is available. The different nitrogen application rates, fertilization depths, and inhibitor types had no significant impact on the EEF and DF yield increase, indicating that under most field management measures, EEF and DF can achieve the goal of increasing crop yield. The yield increase under the EEF condition was directly proportional to the increase in fertilizer application [91,92]. The application of EEF in this study reduced the loss of gaseous nitrogen, ensured that more nitrogen was retained in the soil for crop absorption and utilization, and ultimately improved crop yield. However, our research results further indicated that when the nitrogen application rate was <225 kg ha⁻¹, the yield increase from DF was significantly higher than that from EEF. However, when the nitrogen application rate was high, there was a risk of nitrogen leaching in the DF strategy [31,56], and there was no significant difference in yield between DF and EEF. We must point out that different fertilization depths significantly affect the NUE. Nitrogen is lost in the form of gaseous nitrogen when applied close to the soil surface, and there is also a risk of nitrogen leaching when applied too deep into the soil. Previous research has suggested that the depth of fertilization should not exceed 15 cm to better utilize fertilizers and to increase crop yields. Therefore, this study obtained the highest NUE when the fertilization depth was 5 cm < FD < 15 cm. However, recent research findings suggest that fertilization at a depth of 25 cm may achieve the highest yield and NUE [38,56]. Therefore, from the perspective of overall production and NUE, DF is more beneficial than EEF under different management practices.

The loss of gaseous nitrogen significantly impacts different agricultural management practices. It is worth noting that when the EEF type is a nitrification inhibitor, it may increase NH₃ volatilization, mainly because NIs selectively inhibit the activity of soil nitrifying microorganisms, slowing down the conversion of ammonium nitrogen to nitrate nitrogen in the soil. However, ammonium nitrogen is the substrate for NH₃ volatilization, and the higher the amount of ammonium nitrogen in the soil, the greater the nitrogen loss, which is mainly manifested as NH₃ volatilization. Therefore, a more critical fertilizer management strategy in the future is to combine NIs and UIs [91]. In addition, we found that DF significantly reduced NH3 emissions, and the deeper the fertilization depth, the more the NH₃ emissions were reduced due to the downward transfer of ammonium nitrogen in the soil [38]. Therefore, overall, DF is more beneficial than EEF in reducing NH₃ and N₂O emissions, and DF’s emission-reduction potential is higher than that of EEF.

4.4. Limitations and Implications

Our research results indicate that both DF and EEF can achieve a win–win goal regarding economic and environmental benefits. However, DF increases production and reduces emissions more significantly compared with EEF, so it may be more conducive to sustainable global agricultural ecosystem development. EEF mainly includes NIs (nitrification inhibitors), UIs (urease inhibitors), and coated urea. These three inhibitors are all chemical reagents with complex production processes and high costs. Although specific yield and emission-reduction effects have been achieved under EEF, farmers rarely use this technology due to its high production costs. Therefore, subsidies are needed to encourage farmers to adopt EEF technology. With the increasing level of global agricultural mechanization, traditional manual spraying has been transformed into mechanical fertilization in most regions. Deep fertilization is very easy to achieve and has excellent potential for adoption. Therefore, the DF strategy has more potential for large-scale promotion than EEF.

5. Conclusions

This study established a global database on EEF and DF, aiming to compare their impact on the economic and ecological benefits of agricultural ecosystems. We concluded that DF is more beneficial than EEF in increasing crop productivity and decreasing gaseous nitrogen emissions. The results indicated that the nitrogen application rate, fertilizer type, and fertilization depth are the most important factors affecting DF and EEF. Both DF and EEF can increase the yield and NUE by reducing NH₃ and N₂O emissions. However, DF has a stronger potential for increasing the yield and reducing NH₃ emissions than EEF. Therefore, the results of this study indicate that DF serves as a good substitute for EEF and can be applied globally to improve agricultural ecosystems’ economic and ecological benefits.