Seasonal Patterns in Yield and Gas Emissions of Greenhouse Tomatoes Under Different Fertilization Levels with Irrigation–Aeration Coupling

Abstract

Optimizing aeration, fertilization, and irrigation is vital for improving greenhouse tomato production while mitigating soil greenhouse gas (GHG) emissions. This study investigated the combined effects of three aeration levels (A1: single Venturi, A2: double Venturi, CK: no aeration), two fertilization rates (F1: 180 kg/ha, F2: 240 kg/ha), and two irrigation levels (I1: 0.8 E_pan, I2: 1.0 E_pan) on tomato yield, CO₂, N₂O, and CH₄ emissions, net GHG emissions, net global warming potential (NGWP), and GHG intensity (GHGI) across Spring–Summer and Autumn–Winter seasons. Results showed that aeration and fertilization significantly increased CO₂ and N₂O emissions but reduced CH₄ emissions. Warmer conditions in Spring–Summer elevated all GHG emissions and yield compared to Autumn–Winter seasons. Tomato yield, net GHG emissions, NGWP, and GHGI were 12.05%, 24.3%, 14.46%, and 2.37% higher, respectively, in Spring–Summer. Combining the Maximal Information Coefficient and TOPSIS models, the optimal practice was A1-F1-I1 in Spring–Summer and A2-F1-I1 in Autumn–Winter seasons. These results provide a theoretical basis for selecting climate-smart management strategies that enhance yield and environmental sustainability in greenhouse tomato systems.

1. Introduction

Safeguarding food security, mitigating global climate change, and tackling environmental deterioration are now three critical challenges faced by nations around the world [1,2,3]. CO₂, N₂O, and CH₄ are key greenhouse gases (GHGs) that have a significant impact on the climate and ecosystems. Their extended atmospheric lifetimes and distinctive radiative properties make them significant contributors to global warming [1]. The increase in GHG emissions leads to enhanced radiative forcing and global warming, driving extreme weather events, such as extreme temperatures, heavy precipitation, and more frequent droughts, posing severe threats to human life [2]. Since 1750, the atmospheric levels of these gases have steadily increased, reaching annual averages of 415.7 ppm for CO₂, 1908 ppb for N₂O, and 334.5 ppb for CH₄ by 2021 [3]. Agri-ecosystems are a major source of GHG emissions and have a significant impact on global climate change [4,5]. Consequently, agriculture emerged as the second largest emitter of CO₂, surpassed only by fossil fuel consumption [6]. In addition, agricultural N₂O and CH₄ emissions form 78% and 53% of global emissions from anthropogenic activities [2]. Therefore, reducing GHG emissions from cultivated soils remains a critical challenge and a pressing issue to be addressed.

Tomatoes are among the most commonly consumed vegetables worldwide [7,8]. They are increasingly cultivated in greenhouses in China during spring and autumn seasons, extending the crop’s growing season and ensuring a continuous supply to markets [9]. In greenhouse systems, managing irrigation and fertilization practices is essential for optimizing yields; however, these practices can also lead to higher greenhouse gas emissions and contribute to global warming [10,11]. Therefore, under the dual imperatives of addressing climate change and promoting green agriculture, a critical challenge lies in how to ensure crop yield while effectively reducing greenhouse gas emissions from protected agriculture and improving resource use efficiency. According to existing studies, optimizing farmland management practices in greenhouses—such as irrigation volume, irrigation methods, and fertilization rates—is essential for reducing greenhouse gas emissions, increasing soil organic matter content, and simultaneously improving crop yields [12,13,14].

Combining irrigation and fertilization enables the simultaneous delivery of water and nutrients, making it increasingly popular among farmers globally, particularly in greenhouse tomato cultivation [15]. Nitrogen (N) fertilizers are widely used to achieve high yields, as they play a vital role in crop productivity by enhancing photosynthetic capacity and sink development [16,17]. Additionally, nitrogen (N) fertilizers influence the turnover of soil carbon (C) and nitrogen (N) [18]. Overuse of inorganic N fertilizers can lead to increased soil N₂O emissions by enhancing nitrification and denitrification processes [19]. While optimal yields often require additional N fertilizer, higher N application rates are associated with increased N₂O emissions [20,21,22]. Studies indicate that N fertilizers are responsible for about 60% of global anthropogenic N₂O emissions [23]. Furthermore, the use of N fertilizers impacts soil organic carbon (SOC) levels, which in turn affects CO₂ and CH₄ emissions [24,25]. A meta-analysis has shown that nitrogen application can enhance SOC storage in agricultural soils, potentially reducing CO₂ emissions [13]. However, N fertilizers have also been associated with increased CO₂ emissions due to soil acidification [26]. However, under protected cultivation, improper management practices, such as excessive fertilization and insufficient aeration, have led to soil degradation, low resource use efficiency, and increased greenhouse gas emissions, posing a serious threat to sustainable agricultural development.

In recent years, aerated irrigation has gained considerable interest as a practical approach to reduce inter-root hypoxia. This technique enhances crop yields, improves water use efficiency, boosts fruit quality, and helps to lower soil greenhouse gas (GHG) emissions [27,28,29,30,31,32]. Aerated irrigation supplies aerated water via underground drip irrigation pipes to the crop root zone, alleviating soil hypoxia or oxygen deficiency and enhancing the uptake of water and nutrients by crops. However, altering soil oxygen content also influences nitrifying and denitrifying bacteria, thus affecting soil N₂O emissions [33]. Despite widespread use of aerated irrigation in crops like tomatoes and cucumbers [34,35,36,37], the combined effects of aerated irrigation, irrigation–fertilization, and GHG emissions have not been thoroughly quantified.

Temperature is another critical environmental factor affecting both crop yields and soil GHG emissions [38]. Studies indicate that rising temperatures significantly boost soil GHG emissions in forests [39], grasslands [40], and agricultural land [33,41,42]. For example, increased soil temperature has been shown to reduce wheat yield and N₂O emissions while elevating CO₂ emissions [43]. However, studies have shown that the combination of elevated soil temperature and nitrogen application does not significantly impact soil CO₂ emissions [12], but soil N₂O emissions increased by 26.2% [42]. These findings suggest that the considerable temperature fluctuations during different seasons in greenhouses necessitate taking temperature into account when evaluating crop yields and greenhouse gas (GHG) emissions. Furthermore, current research on the mutual effects of climatic differences and management measures between different seasons period of greenhouse tomatoes on yield and GHG emissions is limited.

Although a few studies have explored the integration of the Maximal Information Coefficient (MIC) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) in agricultural decision-making, their application in the context of irrigation–fertilization–aeration (IFA) regulation for greenhouse tomato systems remains limited. Most existing evaluation frameworks rely on traditional approaches, such as the Analytic Hierarchy Process (AHP) or entropy weighting methods, which, despite their practicality, often suffer from subjectivity and limited capacity to capture nonlinear relationships. In contrast, MIC effectively captures both linear and nonlinear associations between agricultural management practices and key response variables, including crop yield and greenhouse gas (GHG) emissions, while TOPSIS enables a comprehensive ranking of treatment combinations based on multiple performance metrics. This study represents a novel application of the MIC–TOPSIS integration in greenhouse agriculture, aiming to identify optimal IFA strategies that balance yield enhancement with GHG mitigation. By systematically quantifying the synergies and trade-offs among environmental and agronomic outcomes, our proposed framework provides a scientific and objective basis for developing adaptive and sustainable greenhouse crop management practices.

In conclusion, the integrated management of irrigation, fertilization, and aeration plays a crucial role in influencing greenhouse gas (GHG) emissions during greenhouse tomato cultivation. A well-regulated supply of water, nutrients, and oxygen is essential not only for achieving desirable crop yields but also for promoting the sustainability of agroecosystems. To investigate this, the present study conducted field monitoring during two representative cropping periods—Spring–Summer and Autumn–Winter—within the same year. Venturi injectors were used to aerate irrigation water at two oxygenation levels, combined with two fertilization regimes and three aeration intensities. The main objectives of the study were the following: (1) explore seasonal variations in environmental and soil conditions under different combinations of irrigation, fertilization, and aeration; (2) evaluate the impact of these practices on tomato productivity, GHG emissions, net GHG balance, net global warming potential (NGWP), and greenhouse gas intensity (GHGI); and (3) employ the Maximal Information Coefficient (MIC) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to identify critical soil variables associated with GHG emissions and determine the optimal management strategy that enhances yield while reducing emissions. The findings offer strategic guidance for future studies in greenhouse crop management and provide a theoretical basis for simultaneously improving yield and environmental outcomes in greenhouse tomato production.

2. Materials and Methods

2.1. Experimental Location

The experiment was carried out in a greenhouse belonging to the Key Laboratory of Agricultural Soil and Water Engineering in Arid Zones at Northwest A&F University, spanning from March 2021 to January 2022. This controlled facility offered optimal conditions for investigating the impacts of different irrigation and fertilization strategies on tomato growth. The greenhouse is located at 34°20′ N, 108°24′ E, with an altitude of 521 m. The experimental period was divided into two distinct cropping seasons: Spring–Summer (March to July 2021) and Autumn–Winter (August 2021 to January 2022). The location of the site is illustrated in Figure 1. The soil in the greenhouse was classified as silty clay loam, comprising approximately 26% sand, 33% silt, and 41% clay by mass. Within the top 60 cm of the soil profile, the average field capacity was recorded at 32.1% volumetric water content, and the bulk density was measured at 1.35 g·cm⁻³. Before transplanting the tomato seedlings, the surface soil (0–20 cm) was analyzed, revealing the following nutrient contents: organic matter at 9.51 g·kg⁻¹, total nitrogen at 1.86 g·kg⁻¹, total phosphorus at 1.40 g·kg⁻¹, and total potassium at 20.22 g·kg⁻¹. Additionally, environmental data—including air temperature, relative humidity, and solar radiation—were monitored using an automatic weather station (HOBO logger, onset computer corporation, USA) installed within the greenhouse.

2.2. Experiment Design

In this research, tomato seedlings were planted in the greenhouse test plot on 26 March 2021, for the Spring–Summer seasons, with harvesting activities concluding on 15 July of that year. Similarly, for the Autumn–Winter seasons, tomato seedlings were transplanted on 26 August 2021, with harvesting concluding on 15 January 2022. The Spring–Summer seasons trial officially started on 2 April (7 days after irrigation). The Autumn–Winter seasons tests started on 9 September (14 days after irrigation). In total, the transplanting treatment was carried out directly above the drip head and 11 tomato plants were planted in each plot with 35 cm spacing between each tomato plant.

The experiment included two irrigation levels, designated as I1 and I2, which corresponded to 0.8 E_pan and 1.0 E_pan (where 0.8 and 1.0 represent the pan coefficients) [32]. Additionally, two fertilization levels were implemented: 180 kg/ha (F1) and 240 kg/ha (F2), F2 fertilization level reflects the typical fertilization practice adopted by local farmers. The study also examined three aeration conditions: single Venturi aeration (A1), double Venturi aeration (A2), and a control group with no aeration (CK). A completely randomized block design was employed in this study, comprising ten treatments, including A1F1I1, A1F2I1, A2F1I1, A2F2I1, A1F1I2, A1F2I2, A2F1I2, A2F2I2, CKF1I2, and CKF2I2, with each treatment replicated three times. In total, 30 plots were included in the experiment. Topping was performed when the plants had developed four trusses. Except for irrigation, nitrogen application, and aeration, all treatments received identical field management practices, including crop variety, sowing and harvest time, and pest and disease control. The irrigation amounts were calculated using the following formula [28]:

$$W=A\times E_{pan}\times k_{cp}$$

(1)

where W represents the irrigation water volume (L); A denotes the plot area served by a single drip emitter (m²), which is 0.14 m² (0.35 m × 0.4 m); and K_cp is the pan coefficient, set at 0.8 for I1 and 1.0 for I2, respectively.

Irrigation water was delivered using buckets connected to pumps. The total E_pan for the Spring–Summer seasons was 213.28 mm, while the total E_pan for the Autumn–Winter seasons was 119.9 mm. The aeration process employed a Mazzei Venturi meter (Model 287, produced by Mazzei Injector Company, LLC, USA). The field intakes for A1 and A2 were set at 17% and 34%, respectively, as determined by the exhaust method.

During the experiment, various basal fertilizers were applied manually before transplanting the seedlings. These included a biofertilizer derived from sheep waste (excluding nitrogen, phosphorus, and potassium), calcium superphosphate (with a minimum P₂O₅ mass fraction of 16%) at a rate of 150 kg/ha, and water-soluble potassium sulfate (with at least 52% potassium) at 240 kg/ha. The timing and application rate of nitrogen fertilizer were consistent between the Spring–Summer and Autumn–Winter seasons. Specific details regarding nitrogen fertilizer application are provided in

Table 1. Fertilization scale.

Fertilization Level	Base Fertilization	27 d	54 d	68 d	87 d
F2 (240 kg ha−1)	30%	10%	20%	20%	20%
F1 (180 kg ha−1)	30%	10%	20%	20%	20%

.

2.3. Gas Sampling and Analysis

In this experiment, we employed “static dark box-gas chromatography” to measure the emission fluxes of CO₂, N₂O, and CH₄ from the tomato soil throughout the reproductive period. The closed chambers, constructed from 2 mm thick steel plates, measured 25 cm × 25 cm × 40 cm. To enhance insulation, the outer surface was wrapped in sponge and tin foil. Prior to GHG measurements, the base frame, featuring a 3 cm deep groove, was positioned 5 cm below the soil surface to support the gas collection chamber. During the tomato growing period, gas sampling was conducted every 3 to 6 days, with additional samples taken one to two days after irrigation and more frequent sampling following fertilization events. Samples were taken immediately after confinement of the chamber, with gas sampling times of 10:00, 10:10, 10:20, and 10:30, with 30 mL of gas taken each time and analyzed for concentration on the same day. An electronic thermometer (TA288) placed on top of the chamber measured the internal temperature while gas samples were taken to calculate the emission flux. The concentrations of CO₂, N₂O, and CH₄ were determined using a gas chromatograph (Agilent 7890A GC System, USA) equipped with an electron capture detector (ECD). The oven and detector were maintained at temperatures of 55 °C and 350 °C, respectively. Gas flux measurements were performed using the same chromatographic system from Agilent Technologies.

The gas emission fluxes were computed utilizing Equation (2) [44] with the selection of either linear (Equation (3)) or nonlinear regression (Equation (4)) methods being contingent upon the observed patterns of gas concentration changes within the headspace of the closed chambers.

$$F=\rho h{\displaystyle \frac{273}{273+T}}{\displaystyle \frac{dc}{dt}}$$

(2)

where F represents the gas emission flux (CO₂ and CH₄: mg·m⁻²·h⁻¹; N₂O: μg·m⁻²·h⁻¹); ρ denotes the gas density under standard conditions (g·cm⁻³); h indicates the chamber height (m); and dc/dt signifies the curve slope, determined by the changes in gas concentration (CO₂ and CH₄: mg·m⁻²·h⁻¹; N₂O: μg·m⁻²·h⁻¹) in the chamber headspace over time (t, h).

$$\mathrm{c}=a+b\times t(dc/dt=b)$$

(3)

where a, b, and d are constants and t indicates the average air temperature inside the chamber during the sampling period (°C). Other parameters have the same meaning as above.

$$c=a+b\times t+d\times t^2(dc/dt=b)$$

(4)

All parameters have the same meaning as above.

The cumulative emissions of soil CO₂, N₂O, and CH₄ during the whole life cycle of tomato were calculated by Equation (5):

$$R=\sum _{i=1}^n{\displaystyle \frac{F_i+F_{i+1}}{2}}\times D\times 24$$

(5)

where R represents the cumulative amounts of CO₂, N₂O, and CH₄ (kg·ha⁻¹); i indicates the consecutive sampling intervals; D denotes the number of days between two sampling intervals (d); and n signifies the total number of measurements.

2.4. Soil Sampling

In this study, soil oxygen (O₂) content was measured following the method outlined by [45]. Soil measurements were conducted on the day following each irrigation event. The water-filled pore space (WFPS) in the soil was calculated based on water content measurements taken from depths of 0, 10, and 20 cm. These measurements were obtained using the aluminum box oven-drying method, as detailed in Equation (6).

$$\mathrm{WFPS}={\theta }_m·{\displaystyle \frac{{\rho }_b}{1-{\rho }_b/2.65}}$$

(6)

where θ_m denotes the soil water content (%); ρ_b denotes the soil bulk density, equaling 2.65 g/cm³.

The soil NO₃⁻-N contents were quantified using the subsequent formula:

$$M={\displaystyle \frac{1000·C·V}{W}}$$

(7)

where M denotes the soil NO₃⁻-N content (mg/kg); C denotes the measured NO₃⁻-N in the solution sample (mg/L); V denotes the extract sample volume (0.05 L); W denotes the soil sample weight (5 g).

During gas sampling, soil temperature (T) at soil depths of 0, 5, 10 cm were measured with a bent-tube glass mercury thermometer (Wuqiang Hongxing Instrumentation Factory, Hebei Province) to determine the average temperature of the 0–10 cm soil layer.

2.5. Determination of the Tomato Yield

From each plot, the first and last two plants were excluded, and five uniformly growing plants were selected from the remaining for yield measurement. Each treatment consisted of three plots. Yields were calculated per plant and extrapolated to total yield. The tomatoes collected from these plots were precisely weighed using an electronic balance capable of measuring with an accuracy of 5 g.

2.6. Net GHG, NGWP, and GHGI

The net GHG, NGWP, and GHGI are widely recognized indicators for assessing the comparative capacity of GHGs to impact climate change. The duration of this experiment was short and the change in soil organic carbon was negligible [46]. The net GHG is calculated using the following Equation (8) [1,2]:

$$nGHG=273\times R({\mathrm{N}}_2\mathrm{O})+R({\mathrm{CO}}_2)$$

(8)

where R(N₂O) denotes cumulative N₂O amount (kg·ha⁻¹); R(CO₂) denotes cumulative CO₂ amount (kg·ha⁻¹).

The NGWP is calculated using the following Equation (9) [47]:

$$\mathrm{NGWP}=273\times R({\mathrm{N}}_2\mathrm{O})+27.9\times R({\mathrm{CH}}_4)$$

(9)

where R(N₂O) denotes cumulative N₂O amount (kg·ha⁻¹); R(CH₄) denotes cumulative CH₄ amount (kg·ha⁻¹).

The GHGI is calculated using the following Equation (10):

$$\mathrm{GHGI}={\displaystyle \frac{\mathrm{NGWP}}{Y}}$$

(10)

where Y denotes the crop yield (t·hm⁻²).

2.7. Statistical Analysis

Evaluation of the importance of feature parameters. To assess the degree of correlation between feature parameters, such as soil temperature, soil WFPS, soil O₂ content, soil NO₃⁻-N content, and soil CO₂ emissions, N₂O emissions, and CH₄ emissions, thereby determining the importance of these feature parameters in influencing soil CO₂, N₂O, and CH₄ emissions. In this study, the Maximal Information Coefficient (MIC) was selected to measure the degree of linearity or nonlinearity between soil environmental indicators and soil GHG emissions. MIC is used to measure the degree of correlation between characteristic parameters and response variables. Compared to correlation coefficients, MIC values can effectively reflect the nonlinear relationship between characteristic parameters and response variables. The larger the MIC value, the higher the importance of the characteristic parameter to the response variable [48]. In this study, the MINE function package in Python 3.0 software was used to calculate the MIC values between the selected characteristic parameters and soil GHG emissions of CO₂, N₂O, and CH₄. The MIC values were then used to evaluate the importance of the selected characteristic parameters, including soil temperature, soil WFPS, soil NO₃⁻-N content, and soil O₂ content, in influencing the variations of GHG emissions.

The calculation methods and formulas of the overall assessment indices used in the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method are provided in References [32,49].

Excel was utilized for organizing and conducting preliminary analyses of the experimental data. For more comprehensive statistical evaluation, SPSS 26.0 was employed for significance testing and analysis of variance (ANOVA) on the experimental results. The least-significant difference (LSD) method, with a significance level set at p < 0.05 *, was used to compare treatment effects. Furthermore, Pearson correlation analysis was performed using Origin 2019 to explore the relationships between gas fluxes and various soil parameters.

3. Results

3.1. Environmental Variables

During the Spring–Summer seasons, the daily solar radiation at the study site varied between 5.62 W/m² and 147.21 W/m², with an average intensity of 63.72 W/m², whereas during the Autumn–Winter seasons, it ranged from 3.68 W/m² to 87.84 W/m², with an average intensity of 34.19 W/m². On average, solar radiation during Spring–Summer was 1.87 times higher than in Autumn–Winter (Figure 2a,b). This seasonal variation aligns with the 7-year average (2015–2022) recorded at the Yangling weather station.

Daily average air temperatures during the Spring–Summer seasons increased over time, ranging from 14.26 °C to 32.27 °C, with an average intensity of 24.01 °C. In contrast, during the Autumn–Winter seasons, temperatures generally decreased throughout the season, ranging from 29 °C to 7.46 °C, with an average intensity of 16.75 °C. On average, the air temperature during Spring–Summer was 7.26 °C higher than in Autumn–Winter (Figure 2c,d).

The trends in daily average air humidity in the greenhouse were similar for both seasons, ranging from 40.97% to 94.27% during Spring–Summer, with an average intensity of 62.16% and from 49.17% to 96.31% in Autumn–Winter seasons, with an average intensity of 76.16%. The average air humidity during Autumn–Winter was 14% higher than that of Spring–Summer seasons (Figure 2e,f).

3.2. Soil Variables

Soil water-filled porosity (WFPS). The water-filled pore space (WFPS) exhibited a fluctuating downward trend in Spring–Summer and Autumn–Winter seasons, ranging from 31.35–54.02% and 35.59–53.89%, respectively (Figure 3a,b). In addition, higher soil temperature and air temperature in Spring–Summer seasons promoted soil evaporation, resulting in a 3.95% decrease in WFPS compared to the Autumn–Winter seasons (

Table 2. Mean soil temperature, soil water–filled pore space (WFPS), soil NO₃⁻-N content, soil O₂ content, cumulative CO₂ emissions, cumulative N₂O emissions, and cumulative CH₄ emissions of different irrigation, fertilization, and aeration treatments during the Spring–Summer and Autumn–Winter tomato cultivation period.

Season	Treatments		Soil Temperature (°C)	Soil WFPS (%)	Soil NO3−-N Content (mg/kg)	Soil O2 Content (mL/L)	Cumulative CO2 Emissions (kg·hm−2)	Cumulative N2O Emissions (kg·hm−2)	Cumulative CH4 Emissions (kg·hm−2)
Spring – Summer	I	I1	24.47 a	42.27 c	320.04 b	6.78 b	1.22 × 104 b	1.41 ab	−0.74 c
		I2	24.73 a	44.54 b	283.47 c	6.34 c	1.47 × 104 a	1.49 a	−0.45 a
	F	F1	24.58 a	44.31 b	258.89 d	6.59 c	1.31 × 104 b	1.26 c	−0.47 a
		F2	24.67 a	42.95 c	337.30 b	6.45 c	1.42 × 104 a	1.67 a	−0.66 b
	A	A1	24.4 b	43.81 c	291.26 c	6.51 c	1.23 × 104 c	1.22 c	−0.57 c
		A2	24.53 ab	42.6 d	322.79 b	6.83 b	1.51 × 104 a	1.84 a	−0.72 d
		CK	25.27 a	45.35 b	262.38 d	5.89 d	1.35 × 104 b	1.17 c	−0.26 a
Mean	/		24.66	43.69	296.59	6.48	1.36 × 104	1.44	−0.55
Autumn – Winter	I	I1	18.85 b	44.25 b	353.84 a	7.22 a	1.04 × 104 d	1.24 c	−0.84 d
		I2	18.96 b	46.21 a	314.84 b	6.84 b	1.13 × 104 c	1.34 b	−0.68 b
	F	F1	18.68 b	45.55 a	292.59 c	7.13 a	1.06 × 104 d	1.11 d	−0.70 b
		F2	19.15 b	45.30 a	368.28 a	6.85 b	1.14 × 104 c	1.49 b	−0.79 c
	A	A1	18.70 d	45.76 b	323.03 b	6.86 b	1.08 × 104 f	1.18 c	−0.74 d
		A2	18.77 d	44.01 c	357.31 a	7.42 a	1.12 × 104 d	1.58 b	−0.86 e
		CK	19.63 c	47.60 a	291.51 c	6.38 c	1.09 × 104 e	1.00 d	−0.51 b
Mean	/		18.96	45.53	328.77	6.96	1.09 × 104	1.28	−0.73

Note: The different lowercase letters in the same treatment, such as I, F, or A treatment, indicate significant differences (p < 0.05 *) and (p < 0.01 **) between study seasons. I represent I1 and I2 irrigation treatments; F represent F1 and F2 fertilization; A represent A1, A2, and CK aeration treatment, respectively.

). The WFPS values under the high fertilization treatments were generally smaller than the low fertilization treatments, with the F2 treatment reducing by 3.07% over the F1 treatment in Spring–Summer seasons, and by 0.55% in Autumn–Winter seasons (Table 2).

At the same fertilization level, WFPS significantly decreased as aeration levels increased (p < 0.01 **) (Table 2). Among them, A2 treatment was significantly reduced by 2.76% and 6.07% compared to A1 and CK treatments in Spring–Summer seasons, by 3.82% and 7.55% in Autumn–Winter seasons, respectively (p < 0.01 **) (Table 2). However, due to frequent irrigation in the greenhouse, WFPS remained consistently high throughout the experiment (Figure 3a,b).

Soil Oxygen concentration (soil O₂ content). The soil O₂ content during the Spring–Summer period exhibited an initial increase followed by a decrease, ranging from 4.58 mL/L to 8.16 mL/L. In contrast, the Autumn–Winter period saw a steady increase in soil O₂ content, which ranged from 5.12 mL/L to 8.58 mL/L (Figure 3c,d). In addition, higher irrigation levels tended to reduce soil O₂ content during the tomato growing season. The soil O₂ content of I2 treatments were significantly lower than I1 treatments, with the I2 treatment being 6.47% lower than I1 in the Spring–Summer seasons and 5.26% in Autumn–Winter (Table 2). However, the soil O₂ content had a significant increase as the aeration level increased; among them, the A2 treatments were higher by 4.93% and 16.02% than A1 and CK treatments in the Spring–Summer seasons, and 8.25% and 16.43% higher in A2 treatments than A1 and CK treatments in the Autumn–Winter seasons, respectively (p < 0.01 **) (Table 2).

Soil nitrate N content (soil NO₃⁻-N content). Soil NO₃⁻-N content showed a fluctuating downward trend in both the Spring–Summer seasons as well as in the Autumn–Winter seasons, ranging from 95.27 mg/kg to 507.9 mg/kg and from 127.58 mg/kg to 534.36 mg/kg (Figure 3e,f). Notably, four peaks in soil NO₃⁻-N content coincided with the four additional fertilizer applications (Figure 3e,f). The planting season significantly influenced soil NO₃⁻-N content, which was 9.79% lower during the Spring–Summer seasons compared to the Autumn–Winter seasons (p < 0.01 **). The increase in the length of tomato growth due to the air temperature in the Autumn–Winter seasons, caused the soil NO₃⁻-N content to be in a slowly decreasing trend in the later part of the Autumn–Winter seasons (Figure 3f). Aeration significantly affected soil NO₃⁻-N content, which increased with increasing aeration level by 10.83%, 23.03% for A2 treatment over A1, CK treatments in the Spring–Summer seasons, and by 10.61%, 22.57% in the Autumn–Winter seasons, respectively (Table 2). However, soil NO₃⁻-N content decreased significantly with increasing irrigation level, with I2 decreasing by 11.43% from the I1treatment in the Spring–Summer seasons, and by 3.43% in the Autumn–Winter seasons (p < 0.01 **) (Table 2).

Soil temperature (soil T). The seasonal fluctuation in soil temperature within the top 10 cm layer mirrored the trend observed in air temperature (Figure 3d,e,g,h). The Spring–Summer seasons showed a fluctuating upward trend with a range of 16.6 °C–31.9 °C and the Autumn–Winter seasons showed a fluctuating downward trend with a range of 10.7 °C−31.51 °C (Figure 3g,h). The average seasonal soil T during the Spring–Summer seasons was 5.71 °C higher than the Autumn–Winter seasons.

3.3. Soil CO₂ Fluxes

During the Spring–Summer seasons, soil CO₂ fluxes followed a “V”-shaped trend—initially rising before declining—whereas in the Autumn–Winter seasons, the fluxes exhibited a seasonal variation that closely mirrored changes in soil temperature (Figure 4a,b). Specifically, CO₂ fluxes ranged from 227.94 to 1480.26 mg·m⁻²·h⁻¹ during the Spring–Summer seasons and from 67.75 to 938.92 mg·m⁻²·h⁻¹ in the Autumn–Winter seasons. On average, the mean flux in the Spring–Summer seasons reached 641.78 mg·m⁻²·h⁻¹, which was 77.62% higher than that observed in the Autumn–Winter seasons. Cumulatively, the Spring–Summer seasons recorded 1.37 × 10⁴ kg/ha of CO₂ emissions—an increase of 24.69% (p < 0.05 *) compared to the Autumn–Winter seasons. Furthermore, significant interactive effects were observed among irrigation and fertilization, irrigation and aeration, fertilization and aeration, and the three-way combination of irrigation–fertilization–aeration on CO₂ emissions across both seasons (Table 2).

Soil CO₂ emissions significantly increased with higher levels of fertilization; a trend observed in both seasonal periods. Specifically, soil CO₂ emissions rose by 8.34% in F2 treatments compared to F1 treatments in the Spring–Summer period and by 7.52% in the Autumn–Winter period (p < 0.05 *) (Table 2). Notably, emissions were 24.2% and 25.15% higher for the F1 and F2 treatments in the Spring–Summer period compared to the Autumn–Winter period. Additionally, soil CO₂ emissions increased significantly with higher irrigation levels, with I2 treatments showing a 20.46% increase over I1 in the Spring–Summer period, and an 8.91% increase in the Autumn–Winter period (p < 0.05 *) (Table 2). In this context, soil CO₂ emissions were also higher during the Spring–Summer period, being 29.4% and 16.99% greater for the I1 and I2 treatments, respectively, compared to the Autumn–Winter period. Aeration irrigation further enhanced soil CO₂ emissions, with A2 treatments yielding 1.51 × 10⁴ kg/ha in the Spring–Summer period and 1.12 × 10⁴ kg/ha in the Autumn–Winter period. Notably, soil CO₂ emissions for A2 treatments were significantly higher than those for A1 and CK treatments, by 22.37% and 11.88% in the Spring–Summer seasons, and by 3.66% and 2.93% in the Autumn–Winter period (p < 0.05 *) (Table 2). Additionally, the warmer conditions of the Spring–Summer seasons resulted in soil CO₂ emissions being significantly higher for A2, A1, and CK treatments, with increases of 34.95%, 14.32%, and 24.15%, respectively, compared to the Autumn–Winter seasons (p < 0.01 **) (Table 2).

There was a significant positive correlation between soil CO₂ emissions and soil temperature throughout the experimental period (p < 0.01 **). However, soil CO₂ emissions showed a moderately negative but not significant correlation with WFPS, soil NO₃⁻-N content, and soil O₂ content. In addition, there was a strong correlation between soil CO₂ and soil N₂O emissions (p < 0.01 **) (Figure 5). In order to further investigate the significance of the effects of soil temperature, WFPS, soil NO₃⁻-N content, and soil O₂ content on soil CO₂ emissions, soil temperature was defined as the characterizing variable S1, WFPS as the characterizing variable S2, soil NO₃⁻-N content as the characterizing variable S3, and soil O₂ content as the characterizing variable S4; the maximal information coefficient (MIC) model was used to calculate the MIC values between the characterizing variables and CO₂ emissions. The results showed that the MIC value of the characteristic variable S1 with soil CO₂ emissions was the largest at 0.401 in the Spring–Summer seasons (Figure 6a). The MIC values of the characteristic variables S2, S3, and S4 with soil CO₂ emissions were 0.378, 0.382, and 0.342, respectively, and the order of influence on soil CO₂ emissions was S1 > S3 > S2 > S4 (Figure 6a). The MIC value of the characteristic variable S4 with soil CO₂ emissions was the largest at 0.503 during the Autumn–Winter seasons (Figure 6b). The MIC values of the characteristic variables S1, S2, and S3 with soil CO₂ emissions were 0.438, 0.316, and 0.382, respectively, and the order of influence on soil CO₂ emissions was S4 > S1 > S3 > S2 (Figure 6b).

3.4. Soil N₂O Fluxes

Soil N₂O fluxes exhibited fluctuating trends during both the Spring–Summer and Autumn–Winter seasons, with rates ranging from 14.95 to 279.59 μg·m⁻²·h⁻¹ in Spring–Summer and from 12.11 to 119.93 μg·m⁻²·h⁻¹ in Autumn–Winter (Figure 4c,d). Peak fluxes occurred post-fertilization, gradually declining over time (Figure 4c,d). Cumulative N₂O emissions during the Spring–Summer seasons reached 1.46 kg/ha, which was 12.22% higher than in the Autumn–Winter seasons (Table 2).

Fertilization levels significantly influenced soil N₂O emissions, with emissions increasing with higher fertilization. Specifically, N₂O emissions from the F2 treatment rose by 32.45% over the F1 treatment in Spring–Summer and by 33.98% in Autumn–Winter (p < 0.01 **). In addition, warmer Spring–Summer seasons possessed more soil N₂O emissions, and soil N₂O emissions from F2 and F1 treatments in Spring–Summer seasons increased by 12.96% and 11.66% compared to Autumn–Winter seasons (Table 2). Irrigation levels also affected emissions; N₂O fluxes were generally higher in the I2 treatment compared to I1 (Figure 4a,b). Mean emissions from I2 treatments were 1.49 kg/ha in Spring–Summer and 1.34 kg/ha in Autumn–Winter, significantly exceeding those from I1 by 5.76% and 8.13%, respectively (p < 0.05 *) (Table 2). Moreover, emissions were 13.76% and 11.27% higher for I1 and I2 in Spring–Summer compared to Autumn–Winter (Table 2). Aerated irrigation notably enhanced soil N₂O emissions, with emissions rising significantly as the level of aeration increased—a trend consistently observed during both the Spring–Summer and Autumn–Winter seasons (p < 0.01 **) (Table 2). Specifically, in the Spring–Summer seasons, the A2 treatment led to a 50.75% increase in soil N₂O emissions compared to A1, and a 56.98% rise relative to the CK treatment (p < 0.01 **). Similarly, during the Autumn–Winter seasons, N₂O emissions under the A2 treatment were 33.59% and 57.62% higher than those under the A1 and CK treatments, respectively (p < 0.01 **). Furthermore, when comparing across seasons, A2 treatment in the Spring–Summer period resulted in a 16.97% greater N₂O emission than in the Autumn–Winter period (Table 2). In addition, significant interactive effects were found among irrigation–fertilization, irrigation–aeration, and fertilization–aeration combinations in both seasonal periods, all contributing notably to variations in soil N₂O emissions.

Soil N₂O emissions demonstrated a significant inverse relationship with water-filled pore space (WFPS) over the course of the experimental year (p < 0.05 *), while showing a strong positive association with soil NO₃⁻-N concentrations (p < 0.01 **) (Figure 5). To assess the relative influence of key environmental variables (S1, S2, S3, and S4) on soil N₂O emissions, Maximal Information Coefficient (MIC) values were computed between each variable and the N₂O fluxes. The results showed that the MIC value of the characteristic variable S3 with soil N₂O emissions was the largest at 0.761 in the Spring–Summer seasons (Figure 6c). The MIC values of the characteristic variables S1, S2, and S4 with soil N₂O emissions were 0.343, 0.504, and 0.516, respectively, and the order of influence on soil N₂O emissions was S3 > S4 > S2 > S1 (Figure 6c). The MIC value of the characteristic variable S3 with soil N₂O emissions was the largest at 0.761 during the Autumn–Winter seasons (Figure 6d). The MIC values of the characteristic variables S1, S2, and S4 with soil N₂O emissions were 0.689, 0.711, and 0.369, respectively, and the order of influence on soil N₂O emissions was S3 > S2 > S1 > S4 (Figure 6d).

3.5. Soil CH₄ Fluxes

Soil CH₄ fluxes displayed fluctuating trends in both seasons, predominantly showing negative emissions, indicating that the soil acted as a sink for atmospheric CH₄, particularly during the cold Autumn–Winter seasons (Figure 4e,h). Especially during the cold Autumn–Winter seasons, the soil CH₄ fluxes were all negative (Figure 4f). The fluxes ranged from −0.0843 to 0.0185 mg·m⁻²·h⁻¹ in Spring–Summer and from −0.0530 to −0.0055 mg·m⁻²·h⁻¹ in Autumn–Winter (Figure 4e,f). The Autumn–Winter seasons was more conducive to CH₄ uptake, which was reduced by 23.11% during Spring–Summer (p < 0.05 *) (Table 2). Mean soil CH₄ emissions were −0.57 kg/ha in Spring–Summer and −0.74 kg/ha in Autumn–Winter (Table 2). Moreover, in both the Spring–Summer and Autumn–Winter seasons, the interactions among irrigation–fertilization, and fertilization–aeration combinations had highly significant effects on soil CH₄ emissions.

Increased fertilization levels led to higher CH₄ uptake, significantly enhancing absorption in the F2 treatment by 39.5% over F1 in Spring–Summer and by 12.5% in Autumn–Winter (p < 0.01 **) (Table 2). Additionally, F2 and F1 treatments absorbed 16.29% and 32.49% less in Spring–Summer compared to Autumn–Winter, respectively (Table 2). In addition, increased irrigation level significantly promoted soil CH₄ emissions and reduced soil CH₄ uptake. The I2 treatment increased soil CH₄ emissions by 39.19% compared to the I1 treatment during the Spring–Summer seasons, and by 19.81% during the Autumn–Winter seasons (p < 0.01 **). Low air temperatures promoted soil uptake of CH₄ and reduced soil CH₄ emissions. The I2 and I1 treatments increased soil CH₄ emissions by 33.53% and 12.34% throughout the Spring–Summer seasons compared to the Autumn–Winter seasons (Table 2). Aerated irrigation reduced soil CH₄ emissions and emissions decreased significantly with increasing levels of aeration (p < 0.01 **). Soil CH₄ emissions were reduced by 25.87% for the A2 treatment compared to the A1 treatment during the Spring–Summer seasons and by 15.47% during the Autumn–Winter seasons. In addition, soil CH₄ emissions in Spring–Summer were significantly lower than in Autumn–Winter, with reductions of 16.71%, 23.59%, and 69.62% for A2, A1, and CK treatments, respectively (Table 2).

Soil CH₄ emissions correlated positively with soil temperature and WFPS (p < 0.05 *) and negatively with soil NO₃⁻-N content and soil O₂ content (p < 0.001 ***) (Figure 5). A negative correlation also existed between CH₄ and N₂O emissions (p < 0.05 *) (Figure 5). The importance of the influence of the characteristic variables S1, S2, S3, and S4 on soil CH₄ emissions was investigated and the MIC values between the characteristic variables and soil CH₄ emissions were calculated. The results showed that the MIC value of the characteristic variable S2 with soil CH₄ emissions was the largest at 0.918 in the Spring–Summer seasons (Figure 6e). The MIC values of the characteristic variables S1, S3, and S4 with soil CH₄ emissions were 0.423, 0.666, and 0.698, respectively, and the order of influence on soil CH₄ emissions was S2 > S4 > S3 > S1 (Figure 6e). The MIC value of the characteristic variable S2 with soil CH₄ emissions was the largest at 0.918 during the Autumn–Winter seasons (Figure 6f). The MIC values of the characteristic variables S1, S3, and S4 with soil CH₄ emissions were 0.474, 0.670, and 0.504, respectively, and the order of influence on soil CH₄ emissions was S2 > S3 > S4 > S1 (Figure 6f).

3.6. Yield, Net GHG, NGWP, and GHGI

The range of tomato yield in the warm Spring–Summer seasons was 40.37 t/ha–55.52 t/ha, and the range of tomato yield in the Autumn–Winter seasons was 35.97 t/ha–48.13 t/ha. The mean yield of tomatoes was 4.99 t/ha (12.05%) higher in the Spring–Summer seasons than in the Autumn–Winter seasons (

Table 3. Effect of different irrigation, fertilization, and aeration treatments on tomato yield, greenhouse gas intensity (GHGI), net GHG emissions (NGHG), and net global warming potential (NGWP) during the Spring–Summer seasons and Autumn–Winter seasons of tomato plants in greenhouse.

Treatments	Spring–Summer				Autumn–Winter
	Yield (t/ha)	GHGI (kg/ha)	net GHG (kg/ha)	NGWP (kg/ha)	Yield (t/ha)	GHGI (kg/ha)	net GHG (kg/ha)	NGWP (kg/ha)
CKF1I2	40.37 g	6.68 f	13,224.98 h	269.59 h	35.97 g	6.44 fg	10,708.04 h	231.78 f
CKF2I2	44.96 e	7.94 d	14,396.33 e	356.90 e	38.68 f	7.38 de	11,569.04 d	285.48 de
A1F1I1	40.08 g	6.32 g	10,393.18 j	253.41 h	38.49 f	6.18 g	10,124.19 j	237.97 f
A1F2I1	44.85 e	7.32 e	11,144.41 i	328.32 f	40.08 e	7.88 cd	10,834.35 g	315.75 d
A1F1I2	46.59 d	6.29 g	14,068.02 f	293.13 g	40.14 e	6.81 f	11,253.21 f	273.40 e
A1 F2I2	53.30 b	7.44 e	15,065.69 c	396.79 d	41.37 d	9.13 b	12,232.35 b	377.84 c
A2F1I1	41.86 f	8.85 c	13,674.98 g	370.64 e	40.77 de	7.33 e	10,547.37 i	298.68 de
A2F2I1	46.52 d	10.90 a	15,027.46 d	507.23 b	44.03 c	9.29 b	11,475.06 e	409.05 b
A2F1I2	50.03 c	9.26 b	15,962.88 b	463.31 c	46.49 b	8.16 c	11,712.52 c	379.65 bc
A2F2I2	55.52 a	10.65 a	17,719.18 a	591.57 a	48.13 a	11.16 a	12,722.46 a	537.39 a
Significance
I	**	*	**	**	**	**	**	**
F	**	**	**	**	**	**	**	**
A	**	**	**	**	**	**	**	**
I × F	*	ns	**	ns	*	**	**	*
I × A	**	ns	**	**	**	ns	**	**
F × A	*	**	**	**	*	**	**	**
I × F × A	ns	*	**	ns	ns	ns	**	ns

Note: Different lowercase letters indicate significant differences between treatments. I, F, and A represent the effects of irrigation, fertilization, and aeration on each index, respectively; × stands for the interaction effect; I × F and I × A are the interaction effect of irrigation and fertilization, irrigation, and aeration on each index, respectively; F × A is the interaction effect of fertilization and aeration on each index; I × F × A is the interaction effect of irrigation, fertilization, and aeration on each index. * and ** are the marked levels of 0.05 and 0.01, “ns” indicates that I, F, and A have no significant effect on the respective indices, respectively.

). Fertilization, irrigation, and aeration all significantly influenced tomato yield in both seasons. Tomato yield increased significantly with an increasing level of fertilization. Tomato yield increased by 11.97% in F2 treatment over F1 treatment in Spring–Summer seasons and by 5.17% in F2 treatment over F1 treatment in the Autumn–Winter seasons (p < 0.01 **). In addition, tomato yield response to fertilization had a significant seasonal pattern, with the F2 and F1 treatments increasing by 15.47% and 8.46%, respectively, in the Spring–Summer seasons compared to the Autumn–Winter seasons (p < 0.01 **) (Table 3). Irrigation levels had a significant impact on tomato yield, with higher irrigation resulting in greater productivity. During the Spring–Summer seasons, the I2 treatment increased yield by 11.86% compared to I1, while in the Autumn–Winter seasons, the increase was 2.34% (p < 0.05 *). In addition, I2 and I1 treatments in Spring–Summer seasons showed 15.94% and 6.08% higher than I2 and I1 treatments in Autumn–Winter seasons, respectively (Table 3). Aeration was effective in increasing tomato yield and significantly increased tomato yield with an increasing level of aeration. Tomato yield was significantly increased by 4.93% and 13.64% in A2 treatment over A1 and CK treatments in Spring–Summer seasons, and A2 treatment significantly increased by 12.08% and 20.17% over A1 and CK treatments in the Autumn–Winter seasons (p < 0.01 **). The A2, A1, and CK treatments during the warm Spring–Summer seasons increased tomato yield by 8.09%, 15.46%, and 14.3% than the A2, A1, and CK treatments during the Autumn–Winter seasons (Table 3). Moreover, in both the Spring–Summer and Autumn–Winter seasons, the interactions among irrigation–fertilization, irrigation–aeration, and fertilization–aeration combinations had highly significant effects on tomato yield.

During the Spring–Summer seasons, the average net GHG emissions reached 1.41 × 10⁴ kg/ha, which was 2749.85 kg/ha (24.3%) higher than that in the Autumn–Winter seasons (p < 0.05 *) (Table 3). Both irrigation and fertilization exerted significant influences on net GHG emissions, with notable interaction effects observed in both seasonal periods (Table 3). Specifically, the I2 irrigation regime led to a 20.01% increase in net GHG emissions compared to I1 during the Spring–Summer seasons, and an 8.88% increase during the Autumn–Winter seasons (p < 0.05 *). Enhanced fertilization levels also significantly elevated net GHG outputs, with F2 treatments yielding an 8.96% and 8.26% increase over F1 during the Spring–Summer and Autumn–Winter seasons, respectively. Aeration treatments exhibited a strong effect on net GHG emissions, which rose substantially with higher aeration levels. Compared to A1 and CK treatments, A2 treatments increased emissions by 23.12% and 12.93% in the Spring–Summer seasons, and by 4.53% and 4.27% in the Autumn–Winter seasons (Table 3). Additionally, Table 3 reveals that the NGWP and GHGI were 14.46% and 2.37% higher, respectively, in the Spring–Summer seasons compared to the Autumn–Winter seasons. Irrigation, fertilization, and aeration were all found to have highly significant effects on both NGWP and GHGI across seasons. NGWP increased with higher levels of fertilization and irrigation: F2 treatment resulted in increases of 32.16% and 35.46% over F1, while I2 treatment led to increases of 8.31% and 10.22% over I1 in the Spring–Summer and Autumn–Winter seasons, respectively (p < 0.05 *). Moreover, aeration irrigation markedly promoted NGWP, with A2 treatments raising it by 51.99% and 54.25% over A1 and CK in the Spring–Summer seasons, and by 34.84% and 57.06% in the Autumn–Winter seasons (Table 3).

Efficiency evaluation models for combined irrigation, fertilization, and aeration treatments were established, incorporating yield, net GHG, NGWP, and GHGI of greenhouse tomatoes. The TOPSIS method calculated the comprehensive evaluation index for the 10 treatments, detailed in

Table 4. The ranking of irrigation schedule calculated using TOPSIS for drip-irrigated tomato grown in greenhouse under different irrigation, fertilization, and aeration.

Treatments	Spring–Summer				Autumn–Winter
	D+	D−	Ci	Rank	D+	D−	Ci	Rank
CKF1I2	0.24	0.70	0.74	3	0.15	0.73	0.83	2
CKF2I2	0.38	0.50	0.57	5	0.32	0.55	0.63	6
A1F1I1	0.11	0.85	0.89	1	0.17	0.65	0.79	3
A1F2I1	0.17	0.71	0.81	2	0.25	0.57	0.70	5
A1F1I2	0.29	0.67	0.70	4	0.25	0.62	0.72	4
A1F2I2	0.42	0.49	0.54	7	0.54	0.31	0.37	9
A2FII1	0.40	0.48	0.54	6	0.07	0.81	0.92	1
A2F2I1	0.68	0.24	0.26	9	0.47	0.36	0.43	8
A2F1I2	0.59	0.28	0.32	8	0.42	0.41	0.50	7
A2F2I2	0.84	0.11	0.11	10	0.81	0.09	0.10	10

Note: D+ and D− represented positive and negative Euclidean distance, respectively; C_i represented comprehensive evaluation index.

. The comprehensive evaluation revealed that, during the Spring–Summer seasons, A2-F2-I2 emerged as the least favorable combined treatment scenario, with Ci values of 0.113, whereas A1-F1-I1 was identified as the most optimal one with Ci values of 0.891. During the Autumn–Winter seasons, the combined treatment scenario A2-F2-I2 was identified as the least favorable, with a Ci value of 0.102, whereas A2-F1-I1 was deemed the most optimal, exhibiting a Ci value of 0.916. This indicates that combining A1, F1, I1 and A2, F1, I1 were optimal for improving the yield and reducing net GHG, NGWP, GHGI in the Spring–Summer seasons and Autumn–Winter seasons, respectively, under mulched drip irrigation.

4. Discussion

4.1. Soil CO₂ Emissions

Soil CO₂ emissions are governed by a combination of environmental and edaphic conditions. Numerous studies, including [39], have reported significant increases in soil CO₂ fluxes under global warming scenarios across diverse ecosystems—a trend consistent with our findings. In this study, soil temperature exhibited a strong positive correlation with CO₂ emissions (p < 0.01 **) (Figure 5). On average, soil CO₂ emissions were 24.69% higher during the Spring–Summer seasons than in the Autumn–Winter seasons (Table 2). This increase corresponded with seasonal differences in air and soil temperatures, which were 7.26 °C and 5.71 °C higher, respectively, during the Spring–Summer period (Table 2; Figure 2c,d). Nevertheless, the seasonal fluctuation in greenhouse soil CO₂ emissions was less pronounced than that observed in forest ecosystems [39]. Interestingly, some studies, such as [12], found no significant temperature-related variation in CO₂ emissions within wheat–soybean rotations. These discrepancies highlight the crop and system specific responses of soil respiration to thermal shifts, stressing the importance of incorporating seasonal dynamics for accurate GHG emission assessments.

In addition, the three greenhouse farm management practices of aeration, fertilization, and irrigation significantly affected the soil CO₂ emissions. Our findings indicated that soil CO₂ emissions were significantly higher (p < 0.05 *) in the A2 treatment compared to A1 and CK treatments, with increases of 22.37%, 3.66% in the Spring–Summer seasons, and 11.88%, 2.93% in the Autumn–Winter seasons. This result aligns with previous studies [28,33,41]. The primary factor influencing soil CO₂ emissions was the soil O₂ content (Figure 6a,b), which increased with higher aeration levels (Figure 3). An increase in soil oxygen availability improved the rhizosphere environment, thereby promoting microbial proliferation and enhancing enzymatic activity [50], both of which contributed to the rise in soil CO₂ emissions. Furthermore, elevated fertilization levels were associated with increased CO₂ emissions, as evidenced by the observed trends in Figure 4a,b. This increase may be attributed to nitrogen application alleviating plant nitrogen deficiency, promoting root growth and above-ground biomass [36,51], and leading to higher nitrogen inputs into the soil. This, in turn, stimulates microbial activity and promotes soil CO₂ emissions [28,33,41]. Furthermore, nitrogen addition increased soil microbial abundance and enzyme activity, stimulating the mineralization of soil organic matter [46]. Increased irrigation levels also facilitated soil CO₂ emissions, as they elevated microbial activity and accelerated the mineralization and decomposition of soil organic carbon [28,41]. In conclusion, aeration, fertilization, and irrigation promoted soil CO₂ emissions of greenhouse tomato, especially during the warm Spring–Summer seasons.

4.2. Soil N₂O Emissions

Soil N₂O emissions primarily originate from nitrification under aerobic conditions and denitrification under anaerobic conditions [46]. In this study, aeration, fertilization, and irrigation exhibited strong regulatory effects on soil N₂O emissions. Notably, N₂O emissions increased significantly with higher fertilization levels (Table 2). The soil NO₃⁻-N showed a negative correlation with WFPS (Figure 5), and the increased level of fertilization depleted water consumption, leading to a decrease in soil water content, which is also consistent with the previous study by [52]. Furthermore, soil NO₃⁻-N content emerged as the most critical factor influencing soil N₂O emissions (Figure 6c,d). A highly significant relationship (p < 0.01 **) was observed between soil NO₃⁻-N content and soil N₂O emissions (Figure 5), consistent with earlier research [4,19,33]. It had been shown that increased NO₃⁻-N content and elevated NO₂⁻-N availability promote N₂O emissions [28]. Increased irrigation levels also significantly increased soil N₂O emissions. Additionally, higher irrigation levels significantly boosted soil N₂O emissions; the emissions from the I2 treatment were 5.76% and 8.13% higher than those from the I1 treatment during the Spring–Summer and Autumn–Winter seasons, respectively (Table 2). Increasing irrigation levels elevated soil water-filled pore space (WFPS) (Figure 3a,b) and soil moisture, which facilitated denitrification processes mediated by nitrifying bacteria [53], thereby promoting soil N₂O emissions. However, this finding contrasts with the results reported by [54] who observed a decline in soil N₂O emissions with increased irrigation levels. This discrepancy may be attributed to differences in soil texture, climatic conditions, or other agricultural management practices. In addition, increasing the soil O₂ content increased soil N₂O emissions, which in this study, were 50.75% and 56.98% higher in the A2 treatment than in the A1, CK treatment in the Spring–Summer seasons, and 33.59% and 57.62% higher in the Autumn–Winter seasons (Table 2); this result is consistent with the findings of [28,33]. In addition, soil CO₂ and N₂O emissions were found to exhibit a strong correlation (p < 0.01 **) in this study (Figure 5). This phenomenon may be attributed to the fact that soil respiration consumes oxygen, thereby accelerating the formation of an anaerobic microenvironment conducive to denitrification, which in turn, enhances soil N₂O production and emission [55]. This is consistent with previous findings indicating that denitrification is a major pathway for N₂O generation under oxygen-limited conditions [56]. Elevated soil temperatures during the warm Spring–Summer seasons also promoted soil N₂O emissions, which increased by 12.22% during the Spring–Summer seasons compared to the Autumn–Winter seasons. In contrast, studies by [57,58] indicated that warmer and drier conditions might slow the decomposition of soil organic carbon and alter the communities of nitrifying and denitrifying bacteria, leading to reduced CO₂ and N₂O emissions from the soil. To mitigate GHG emissions, reducing fertilization levels, aeration, and moderating irrigation practices present viable strategies.

4.3. Soil CH₄ Emissions

The soil CH₄ emissions are primarily in anaerobic environments. Organic acids, CO₂ and H₂ produced by microbial metabolism are used by methanogenic and methane-oxidizing bacteria to produce CH₄. This process predominantly occurs in soils with high moisture content and limited aeration. Numerous studies have confirmed that dryland agricultural soils often act as sinks for CH₄ due to their aerobic conditions and low methanogenic activity [59]. In this study, unlike the consistent positive emissions of CO₂ and N₂O throughout the two growing seasons, soil CH₄ showed no clear variation pattern and functioned as a sink for atmospheric CH₄ fluxes most of the time (Figure 4e,f). This indicates that under the mildly aerobic conditions of the study area, the soil had strong methane oxidation potential and effectively absorbed atmospheric CH₄, contributing to ecological mitigation of methane emissions. Moreover, seasonal variation had no significant effect on CH₄ emissions, but both CH₄ emission and uptake were influenced by irrigation, fertilization, and soil physicochemical properties, with irrigation exerting a particularly significant impact. The WFPS was a key factor influencing soil CH₄ emission (Figure 6e,f), which showed a positive correlation with soil CH₄ emission (p < 0.05 *), and this was consistent with the results of [41]. The reason for this result is due to the fact that under anaerobic conditions, methanogenic bacteria decompose carbonaceous organic matter in the soil into CH₄ [60]. In addition, aeration reduces soil CH₄ emissions, and soil CH₄ emissions decrease significantly with increasing levels of aeration (Table 2). Aerating improves soil aeration, increases soil O₂ content, and methane oxidizing bacteria oxidize CH₄, which in turn, reduces soil CH₄ emissions [41]. Our study corroborated this, revealing a highly significant relationship (p < 0.001 ***) between soil CH₄ emissions and soil O₂ content (Figure 5). Therefore, moderate aeration and lower soil moisture conditions are conducive to enhancing the soil’s CH₄ uptake potential. Fertilization also plays a crucial role in affecting soil physicochemical properties, which subsequently influence soil CH₄ emissions. In our study, higher levels of fertilization were associated with reduced soil CH₄ emissions, consistent with findings from [60]. This reduction may result from increased activity of methane-oxidizing bacteria, which enhances the oxidation of inter-root CH₄ within crops, thereby decreasing overall soil CH₄ emissions. A significant interaction between aeration and fertilization was observed, indicating that CH₄ uptake rates may be enhanced or suppressed under certain coupling levels.

4.4. Impact of the Coupling of Irrigation, Fertilization, and Aeration on Yield, Net GHG, NGWP, and GHGI

The four factors considered in this study—aeration, fertilization, irrigation, and growing season—not only influence the emissions of greenhouse gases (GHGs) such as N₂O, CO₂, and CH₄, but also significantly impact crop yield by modifying key soil environmental conditions, including water availability, nutrient supply, aeration status, and thermal regime [33,38]. In this study, the highest tomato yield was achieved under the treatment combining high levels of aeration, fertilization, and irrigation (A2F2I2), which significantly exceeded the yield of the control treatment (CKF1I2) by 35.7% (p < 0.01 **) (Table 3). This finding aligns with previous studies [32,35,37]. The improved aeration, fertilization, and irrigation practices enhanced the soil environment, increased nutrient availability, and promoted nitrogen uptake by plants, ultimately boosting yield [61]. Furthermore, tomato yield was found to be 12.05% (p < 0.05 *) higher during the Spring–Summer seasons compared to the Autumn–Winter seasons. However, net GHG, NGWP, GHGI also increased by 24.3%, 14.46%, and 2.37% (Table 3). Therefore, for the cultivation of tomatoes in greenhouse systems, finding a balance between yields and GHG emissions to minimize the negative impacts on the environment is the key to this study.

To achieve economic benefits for farmers while simultaneously reducing GHG emissions, effective field management practices are essential. In this study, the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method was employed to identify the optimal irrigation combinations. As a robust multi-criteria decision-making tool, TOPSIS is particularly effective in evaluating the relative advantages and disadvantages of various alternatives. It has been widely applied to the assessment of irrigation strategies under different fertilization and aeration conditions, as well as in evaluating water-saving practices [62]. In our previous research, we also utilized the TOPSIS approach to comprehensively assess the performance of greenhouse tomato cultivation during the Spring–Summer season, enabling the selection of optimal irrigation, fertilization, and aeration combinations that jointly enhanced the yield, fruit quality, and water productivity coefficient (WP_C) [32]. In this study, the TOPSIS method analysis showed that A1F1I1 was the optimal combination for the Spring–Summer seasons, and A2F1I1 was the optimal combination for the Autumn–Winter seasons, and these combinations of experimental were able to balance the increase in greenhouse tomato yield and the reduction in GHG emissions.

5. Conclusions

In this study, we systematically evaluated the effects of different levels of aeration, fertilization, and irrigation on tomato yield, GHG emissions, net GHG emissions, NGWP, and GHGI under greenhouse conditions across Spring–Summer and Autumn–Winter seasons. The results demonstrated the following:

(1)

Seasonal interactions among aeration, fertilization, and irrigation significantly influenced soil GHG emissions. Specifically, higher levels of aeration, fertilization, and irrigation increased CO₂ and N₂O emissions, while enhanced aeration and fertilization reduced CH₄ emissions. Compared to the Autumn–Winter seasons, Spring–Summer CO₂, N₂O, and CH₄ emissions rose by 24.69%, 12.22%, and 23.11%, respectively (p < 0.05 *).

(2)

All three management practices notably improved tomato yields but also intensified GHG-related environmental impacts. In Spring–Summer seasons, tomato yield, net GHG emissions, net global warming potential (NGWP), and greenhouse gas intensity (GHGI) were 12.05%, 24.3%, 14.46%, and 2.37% higher than in Autumn–Winter seasons, respectively (p < 0.05 *).

(3)

Considering both economic and environmental benefits, our analysis, which integrated the MIC and TOPSIS models, identified the A1F1I1 treatment (Aeration level, single aeration (A1); fertilization level, 180 kg/ha (F1); irrigation level, 0.8 E_pan (I1)) as the optimal combination for the Spring–Summer seasons and the A2F1I1 treatment (Aeration level, double aeration (A2); fertilization level, 180 kg/ha (F1); irrigation level, 0.8 E_pan (I1)) as the ideal combination for the Autumn–Winter seasons.

These findings offer a robust scientific foundation and theoretical reference for future research aimed at enhancing greenhouse tomato yields while simultaneously reducing emissions.