Sustainable Irrigation Management of Winter Wheat and Effects on Soil Gas Emissions (N₂O and CH₄) and Enzymatic Activity in the Brazilian Savannah

Abstract

Water scarcity and greenhouse gas (GHG) emissions pose significant challenges to sustainable wheat production in tropical regions such as the Brazilian Cerrado. This study evaluated the effects of different soil water depletion levels, denoted as f (20%, 40%, 60%, and 80% of available water capacity—AWC), on no-tillage winter wheat irrigated after rainfed soybean cultivation. Grain yield decreased significantly at depletion levels ≥ 60%, with the highest yields observed at f = 20% (6933 kg ha⁻¹) and f = 40% (6814 kg ha⁻¹). Water use efficiency (WUE) ranged from 12.4 to 14.0 kg ha⁻¹ mm⁻¹, with no significant differences among treatments. Nitrous oxide (N₂O) emissions peaked at f = 60% (4.55 kg ha⁻¹), resulting in the highest average global warming potential (GWP = 1.185.78 kg CO₂ eq ha⁻¹) and greenhouse gas intensity (GHGI = 192.66 kg CO₂ eq Mg⁻¹ grain). Methane (CH₄) acted as a net sink across all irrigation levels. Soil enzymatic activities (β-glucosidase and arylsulfatase) were not significantly affected by irrigation management. Overall, irrigation scheduling based on f = 40% soil water depletion provided the best balance between productivity and environmental sustainability, representing a climate-smart and resource-efficient strategy for wheat production in tropical agroecosystems. These findings provide promising insights for tropical agriculture by showing that sustainable irrigation can balance productivity and climate mitigation in the Cerrado. Maintaining soil water depletion below 60% significantly reduces N₂O emissions and environmental impact, emphasizing the importance of conservation practices. Additionally, preserving soil biological quality supports the long-term viability of these practices and offers valuable guidance for policies promoting efficient irrigation in climate-vulnerable regions.

1. Introduction

The rising food demand is causing pressure on agriculture towards intensification as well as sustainability of production [1]. In addition, the trend of increasing GHG concentrations is expected to continue in the coming decades. A quarter of total anthropogenic GHG emissions is believed to arise mainly from deforestation, ruminant livestock, and fertilization [2,3]. In recent decades, methane (CH₄) and nitrous oxide (N₂O) emissions from agriculture have increased rapidly, by 265 and 125%, in relation to the pre-industrial (before 1750) levels [4]. In Brazil, research has shown that in 2021, emissions from food systems corresponded to 1.8 billion tons of carbon dioxide equivalent (GtCO₂ eq)—in other words, 73.7% of the total gross emissions in the same year, when 2.4 GtCO₂ eq were emitted nationwide. However, in 2023, the agricultural sector was responsible for the highest net GHG emissions from the national territory (631,176,931 tons of CO₂ equivalent) [5].

In view of the global climate change and increasing water scarcity, improving irrigation management and reducing GHG emissions in agriculture is critical to achieving sustainable food production and climate resilience, particularly in sensitive tropical ecosystems such as the Brazilian Cerrado. This scenario stimulates the transition from conventional production models to climate-smart systems focused on GHG mitigation and adaptation, given the lack of GHG data for irrigated wheat in the Cerrado.

For grain crops, nitrogen fertilization and soil, crop, and irrigation management are major drivers of greenhouse gas (GHG) emissions [6,7]. These practices influence soil water distribution and water use efficiency, thereby affecting GHG [6,8]. Soil wetting and drying cycles can stimulate microbial activity (Birch effect), promoting residue mineralization and denitrification of accumulated nitrate [9]. Additionally, soil enzymes, secreted by soil microorganisms, can influence microbial communities and enzymatic activity, which in turn affects nutrient cycling and crop yield [10]. Most biochemical transformations are mediated by enzymes and influenced by soil management. In other words, soil enzymatic activity is a sensitive indicator of land use change e diferentes níveis de depleção de água no solo [11]. The determination of soil enzymatic activity allows an evaluation of the impact of management practices on soil microbiota and, consequently, on N mineralization and N₂O emissions. However, ref. [12] did not describe the correlation between enzymatic activity and N₂O emissions in an integrated crop–livestock system.

Despite advances in wheat irrigation in the Cerrado, data on the mechanisms of greenhouse gas (GHG) emissions associated with different water management levels remain scarce. Variation in soil water depletion, when used as a criterion for irrigation scheduling, directly affects soil arability and nutrient availability, influencing nitrification and denitrification processes. These microbial transformations are significant sources of nitrous oxide (N₂O) and methane (CH₄) emissions, whose intensity can be modulated by irrigation frequency and nitrogen fertilizer use. Therefore, understanding the mechanisms controlling GHG emissions under different irrigation management practices is essential to establish sustainable and resilient production strategies in the context of the Cerrado [13].

The use of irrigation technologies requires appropriate water management strategies that ensure both economic and environmental viability. Primarily, crop-specific irrigation parameters must be defined, including the optimal timing and application depth, to maximize yield and water use efficiency (WUE) [14]. Monitoring the sustainability of a system is also crucial to ensuring a long-term, rational use of natural resources.

For irrigation, it is assumed that the water available to crops (available soil water capacity (AWC)) is stored in the profile corresponding to the effective depth of the root system and is computed as the difference between soil moisture at field capacity (FC) and the permanent wilting point (PMP) [15]. However, when AWC depletion level drops below a certain limit, crop yields are affected by the soil water condition. Therefore, depending on the crop, climate, and soil conditions, the depletion limit (f) fraction is established as an indicator of the moment of irrigation [16].

An accurate estimation of crop evapotranspiration (ETc) is essential for an efficient irrigation management. A common approach involves crop coefficients (Kc) combined with reference evapotranspiration (ETo) [17]. The Kc varies with the crop species, growth stage, and local climate and is experimentally determined under ideal, unstressed cultivation conditions [18]. Several studies have addressed the measurement of the crop coefficient (Kc) for wheat in different regions of the world [19,20,21]. However, due to advances in breeding and the evolution of management practices, region-specific studies that consider local soil and climatic conditions are necessary. Generalized Kc values may fail to accurately represent specific regional realities [22].

Wheat (Triticum aestivum L.) accounts for approximately 30% of the world grain production [23]. In Central Brazil, irrigated wheat is currently planted on an area of around 30 thousand hectares [24]. More than 29% of the national grain production is grown in the central-western region of Brazil [25]. Wheat is an option for crop sequence and rotation systems of grain, vegetable and fiber production in that region, where it is grown in part of the cultivation break of soybean. Due to the selected plant material and regional climate, the yield of wheat cultivated in the Cerrado is high (≈6000 kg ha⁻¹) and the industrial quality is excellent [26]. With the significant expansion of irrigated wheat cultivation into the Cerrado region of Central Brazil, references for water use efficiency and GHG emissions must be established that contribute to sustainable production systems.

The research presents an original and integrated approach by evaluating the effects of irrigation scheduling thresholds (f) on wheat production in the Brazilian Cerrado, considering not only crop performance and water use efficiency but also greenhouse gas emissions and soil enzymatic activity related to nutrient cycling. This combination is seldom explored in the literature and is essential for developing climate-smart agricultural strategies tailored to tropical conditions.

In this context, the objective of the study was to evaluate the effects of different levels of soil water depletion (f) as an indicator for irrigation scheduling during the development of no-tillage winter wheat, following rainfed soybean. In addition, the study evaluated the effects on grain yield, greenhouse gas emissions (N₂O and CH₄), water use efficiency (WUE), and soil enzymatic activity and determined the crop coefficient (Kc) at soil water depletion levels when irrigation should be resumed. This study hypothesizes that soil water depletion levels used as criteria for irrigation management significantly affect wheat grain yield, water use efficiency (WUE), greenhouse gas emissions (N₂O and CH₄), and soil enzymatic activity. Lower depletion levels (i.e., more frequent irrigation) are expected to improve crop performance but may also enhance microbial activity, leading to increased GHG emissions.

2. Materials and Methods

2.1. Experimental Site

The study was carried out in an experimental area of Embrapa Cerrados (Figure 1), in Planaltina, Distrito Federal, Brazil (15°33′33.99″ S, 47°44′12.32″ W; 1035 m asl), from May 2022 to March 2024. According to the Köppen–Geiger classification, the climate in the study area is Aw, i.e., tropical with dry winters and mean temperature above 18 °C in the coldest month [27]. The mean annual precipitation is 1394 mm, of which 87% falls between October and April [17]. During the first crop year (2022–2023), total precipitation amounted to 31.8 mm from May to September 2022 and 987.6 mm from October 2022 to April 2023, with mean temperatures of 20.64 °C and 21.81 °C, respectively. In the second crop year (2023–2024), total precipitation was 53.4 mm from May to September 2023 and 964.8 mm from October 2023 to April 2024, with mean temperatures of 21.58 °C and 23.44 °C, respectively. The soil was classified as Red Yellow Latosol [28], with specific physical–hydraulic (

Table 1. Soil physical–hydraulic properties in the experimental area.

Soil Depth	Sand	Clay	Silt	BD	FC	PWP	AWC
(cm)	(%)	(%)	(%)	(g cm−3)	(cm3 cm−3)	(cm3 cm−3)	(mm)
00–20	33	52	16	1.08	0.32	0.19	26.00
20–40	32	54	14	1.06	0.31	0.19	24.40

BD = soil bulk density; FC = field capacity; PWP = permanent wilting point, and AWC = available soil water capacity.

) and chemical properties (

Table 2. Soil chemical properties in the experimental area.

pH	OM	P	K+	Ca2+	Mg2+	Al3+	H + Al	CEC
-	(g kg−1)	(mg dm−1)		(cmolc kg−1)
5.27	23.80	21.76	223.23	4.32	1.08	0.11	3.89	9.88

OM = organic matter; CEC = cation exchange capacity.

).

2.2. Experimental Design and Cultural Practices

The study was carried out from May 2022 to March 2024. Growing seasons 1 and 2 began in May 2022 and 2023, respectively, with wheat planting, and ended in March 2023 and 2024, respectively, with the soybean harvest. The experiment performed under no till, consisting of twelve 6 × 6 m plots, separated by 12 m interrows. The experiment was arranged in a randomized block design, with three replications. Wheat (cv. BRS 264) was sown on 23 May 2022 and 22 May 2023. The amount of planted seed was calculated to ensure a planting density of 420 healthy plants per square meter. Each plot consisted of 28 plant rows, spaced 0.175 m apart. For fertilization, 400 kg ha⁻¹ of the formula 4-30-16 (N-P₂O₅-K₂O) was applied. Urea nitrogen was applied as top dressing, at a rate of 120 kg ha⁻¹, 15 days after soybean emerged.

Soybean (cultivar BRS 7482 RR) was sown on 6 November 2021, 8 November 2022, and 1 November 2023. The amount of seed was calculated to ensure a planting density of 35 healthy plants per square meter. Each plot consisted of 12 plant rows, spaced 0.5 m apart, and was fertilized with 600 kg ha⁻¹ of the formula 0-20-20 (P₂O₅-K₂O) at planting. Soybean was cultivated exclusively to represent the soybean–wheat production system, with rainfall serving as the sole water source during the growing season.

2.3. Wheat Water Management, Yield, Water Use Efficiency, and Crop Coefficient

The treatments were defined according to the depletion of the available water capacity (f) in the soil, by applying irrigation when AWC was reduced by 20%, 40%, 60%, and 80%, due to crop evapotranspiration. Thus, the treatments were identified as: 20%, 40%, 60%, and 80%. In this study, the effective depth of the root system, defined at 0.4 m for wheat, was taken into consideration for the AWC [29]. The irrigation system was described in detail by [15].

After planting, all plots were irrigated with 20 mm water, at 2-day intervals, until soil moisture at the effective root depth was raised to field capacity (FC). After plant germination, soil moisture was monitored daily with a neutron probe (CPN 503 Elite Hydroprobe^®, CPN International, a division of InstroTek, Inc., Concord, CA, USA). In the center of each plot, aluminum tubes were installed to a depth of 0.70 m to access the probe. Soil moisture readings were taken between 8:00 and 9:00 a.m., at depths of 0.10, 0.30, and 0.50 m. Irrigation was applied whenever moisture monitoring confirmed the AWC depletion level for each treatment [15] (Figure 2). The effects of the different treatments on wheat yield (Y) were evaluated for each plot by harvesting five 6 m rows. Yield was estimated after correcting grain moisture to 13%. Water use efficiency (WUE) was calculated as the ratio between crop yield (Y) and cumulative evapotranspiration in the crop cycle (ETc) [15].

The crop coefficient (Kc) was estimated from the treatment in which grain yield (Y) reached the highest statistically significant value, indicating that no water stress occurred during winter wheat development. The coefficient was calculated using the following standard equation [30]:

$$\mathrm{K}\mathrm{c}={\displaystyle \frac{\mathrm{E}\mathrm{T}\mathrm{c}}{\mathrm{E}\mathrm{T}\mathrm{o}}}$$

Crop evapotranspiration (ETc) was estimated using the soil water balance model based on the following equation [31]:

$$\mathrm{E}\mathrm{T}\mathrm{c}=\mathrm{Pr}+\mathrm{I}+\mathrm{Q}+\mathrm{\Delta }\mathrm{A},$$

where Pr is rainfall (mm); I is the irrigation depth applied (mm); Q is the water flux leaving (drainage) or entering (capillary rise) the soil profile (mm); and ΔA is the change in soil water storage within the monitored profile (mm). Water balance was calculated for the periods between irrigations, and in the absence of rainfall, negligible capillary rise was assumed. Thus, the soil water balance equation was simplified to

ETc = ∆A

Soil moisture values were used as a function of depth (θ(Z)) to estimate water storage in the soil profile based on the trapezoidal rule, according to the following equation [32]:

$$\mathrm{h}\mathrm{z}=\left[\mathrm{\theta }\right(\mathrm{Z}10)+\mathrm{\theta }(\mathrm{Z}30)+0.5(\mathrm{\theta }\left(\mathrm{Z}50\right)\left)\right]∆\mathrm{Z},$$

where hz is the water storage in the 0–50 cm soil layer; θ(Z10), θ(Z30), and 0.5 × θ(Z50) represent the volumetric water content at soil depths of 0–20, 20–40, and 40–50 cm, respectively; and ΔZ is the thickness of the soil layer between measurement points, defined at 20 cm.

Initial water storage (hzi) was estimated on the morning following irrigation, at approximately 9:00 a.m. Irrigations were scheduled to end by 5:00 p.m. to allow for nighttime water redistribution within the soil profile. Final water storage (hzf) was estimated immediately before the next irrigation event.

Water storage change between irrigation events was calculated as the difference between post-irrigation and pre-irrigation soil water content in the control profile, according to the following equation:

∆A = hzi − hzf

Reference evapotranspiration (ETo) was estimated by the Penman–Monteith equation [18]:

$$\mathrm{E}\mathrm{T}\mathrm{o}={\displaystyle \frac{0.408·\mathrm{\Delta }·\left(\mathrm{R}\mathrm{n}-\mathrm{G}\right)+\mathrm{{\rm Y}}·\frac{900}{\mathrm{T}+273}·{\mathrm{U}}_2·({\mathrm{e}}_{\mathrm{s}}-{\mathrm{e}}_{\mathrm{a}})}{\mathrm{\Delta }+\mathrm{{\rm Y}}·(1+0.34·{\mathrm{U}}_2)}},$$

where ETo (mm day⁻¹), ∆ is the slope of the saturation vapor pressure curve (kPa °C⁻¹); Rn is the net radiation at the crop surface (MJ m⁻² day⁻¹); G is the soil heat flux density (MJ m⁻² day⁻¹); γ is the psychrometric constant (kPa °C⁻¹); T is the mean daily air temperature at 2 m height (°C); U₂ is wind speed at 2 m height (m s⁻¹); e_s is the saturation vapor pressure (kPa); and e_a is the actual vapor pressure (kPa). The climatic variables were recorded by an automatic weather station (Campbell Scientific) close to the experimental area.

2.4. GHG Sampling, Environmental Variables and Soil Enzymes

To quantify CH₄ and N₂O concentrations between soil and atmosphere from 25 May 2022 to 15 March 2024, the field method of closed static chambers was used. The protocol and materials for gas sampling, emission calculation, and identification of covariates related to soil GHG emissions were described by [33].

Fluxes were continuously sampled 127 times. The sampling frequency of GHG fluxes was appropriate to adequately represent and measure the differences between soil water depletion levels. Sampling intervals were defined by events such as input application, rainfall, irrigation, planting date, N fertilization, harvest date, senescence, and others.

In each experimental plot, two static chambers were installed in the planting row for the duration of the wheat growth cycle. Later, 25 days after soybean had emerged, the chambers were placed in between the plant rows. Gas samples were taken between 9:00 and 11:00 a.m., as recommended by [7]. Each chamber was sampled at 0, 15, and 30 min following closure. Thermometers were positioned within each chamber to monitor the internal air temperature. At each sampling, the gas was sucked out of the chamber in a 60 mL syringe with a three-way valve and injected into a pre-evacuated 25 mL vacuum vial. Additionally, one sample per block was collected as a reference for the atmospheric air standard. Soil temperature was determined at each sampling event using a digital thermometer, positioned 5 cm below the surface. Gas concentrations were determined by gas chromatography (Nexis GC-2030, Shimadzu, Kyoto, Japan), with a capillary column 30 m × 0.53 mm × 20 µm Rt-QS-BOND (Restek, Bellefonte, PA, USA)). N₂O concentrations were measured by an electron capture detector (ECD), while CH₄ and CO₂ by flame ionization detector (FID) (Shimadzu, Kyoto, Japan). The gas chromatograph was calibrated using certified gas standards (Messer, Bad Soden, Germany), consisting of a special gas mixture diluted in high-purity N₂. Calibration was performed with the following concentration ranges for N₂O 0.2 (±0.0), 0.5 (±0.001), 1.0 (±0.1), and 2.0 (±0.1) µmol mol⁻¹; CH₄ 1.0 (±0.0), 10 (±0.1), 50 (±0.1), and 100 (±0.1) µmol mol⁻¹; and CO₂ 200 (±0.1), 500 (±0.5), 1000 (±0.8), and 2000 (±2.5) µmol mol⁻¹. All calibration curves exhibited coefficients of determination (R²) greater than 0.99. The estimated quantification limits ranged from 0.08 to 0.13 µmol mol⁻¹ for N₂O, from 0.51 to 2.55 µmol mol⁻¹ for CH₄, and from 0.60 to 2.88 µmol mol⁻¹ for CO₂. The corresponding detection limits ranged from 0.03 to 0.04 µmol mol⁻¹ for N₂O, from 0.17 to 0.84 µmol mol⁻¹ for CH₄, and from 0.20 to 0.95 µmol mol⁻¹ for CO₂. The fluxes were then determined as a function of the sampling time [34], and cumulative emissions were estimated by linear interpolation [33].

The global warming potential (GWP) of GHG emissions, based on a 100-year horizon and Intergovernmental Panel on Climate Change (IPCC) guidelines, was estimated by the following equation:

GWP (kg CO₂ eq. ha⁻¹) = 265. × cumulative_N₂O (kg ha⁻¹) + 28. × cumulative_ CH₄ (kg ha⁻¹).

The greenhouse gas emission intensity (GHGI) was then calculated according to [7] by this equation:

$$\mathrm{G}\mathrm{H}\mathrm{G}\mathrm{I}\left(\mathrm{kg }\mathrm{C}{\mathrm{O}}_2\mathrm{ e}\mathrm{q}.\mathrm{M}\mathrm{g }\mathrm{M}{\mathrm{g}}^{-1}\mathrm{ g}\mathrm{r}\mathrm{a}\mathrm{i}\mathrm{n}\right)={\displaystyle \frac{\mathrm{G}\mathrm{W}\mathrm{P}}{\mathrm{G}\mathrm{r}\mathrm{a}\mathrm{i}\mathrm{n }\mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}}}$$

During gas sampling, soil samples were systematically collected from a depth of 0 to 10 cm to determine mineral N (NO₃⁻ and NH₄⁺). Two subsamples per experimental plot were collected to form a composite sample that was analyzed by automatic flow analysis [35,36]. Soil bulk density was determined by the volumetric cylinder method, and particle density was considered to be 2.65 g cm⁻³ [37]. Soil moisture was determined by the gravimetric method and WFPS by the following equation:

$$\mathrm{W}\mathrm{F}\mathrm{P}\mathrm{S}={\displaystyle \frac{\mathrm{\theta }·\frac{\mathrm{B}\mathrm{D}}{\mathrm{W}\mathrm{D}}}{1-\left(\frac{\mathrm{B}\mathrm{D}}{\mathrm{P}\mathrm{D}}\right)}}$$

where WFPS is the water-filled pore space (%); θ—gravimetric water content (g g⁻¹); BD—soil bulk density (g cm⁻³); WD—water density (1.0 g cm⁻³); and PD—particle density.

The activities of arylsulfatase and β-glucosidase were determined using a colorimetric method based on the quantification of p-nitrophenol (PNP), released after enzymatic hydrolysis of specific substrates, as originally described by [38] and widely adapted in recent studies [39]. Soil samples were air-dried, passed through a 2 mm mesh sieve, and incubated with buffered substrate solutions at 37 °C for 1 h. For β-glucosidase activity, p-nitrophenyl-β-D-glucopyranoside was used as the substrate, while arylsulfatase activity was measured using p-nitrophenyl sulfate. Following incubation, the reaction was stopped with 0.5 M CaCl₂ and 0.1 M NaOH, and the absorbance of the released PNP was measured spectrophotometrically using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan) at 410 nm for arylsulfatase and 420 nm for β-glucosidase. Enzyme activities were expressed as mg of PNP released per kg of dry soil per hour (mg PNP kg⁻¹ h⁻¹). This method is widely used to assess microbial functional activity associated with sulfur and carbon cycling in soil ecosystems under different land use and management conditions [39,40].

2.5. Statistical Analysis

All data were checked for normality of residuals by the Shapiro–Wilk test and for homogeneity of variances by Hartley’s maximum F ratio and the Bartlett test. The N₂O and CH₄ flux data were subjected to analysis of variance (ANOVA), and the means compared by the Tukey test (p = 0.05). For comparison, the cumulative fluxes of N₂O and CH₄ were linearly interpolated for the sampling dates, in relation to the soil water depletion levels. Pearson’s correlation analysis was used to determine the relationship between GHG and environmental variables. Statistics were analyzed using the statistical package of R software, version 3.2.2. The Kc data were analyzed by regression, using the model with the highest coefficient of determination (R²) and a significant linear or quadratic effect at the 5% probability level.

3. Results

3.1. Yield, Post-Emergence Irrigation Period, Number of Irrigation Events, Applied Water Depth per Cycle, Water Use Efficiency, and Crop Coefficient

Normality of residuals was confirmed by the Shapiro–Wilk test (α = 0.05). Therefore, analysis of variance (ANOVA) was applied to evaluate the yield data. Combined analysis was performed, since the ratio between the highest and lowest residual mean square from the individual ANOVAS of the experiments conducted in 2021, 2022, and 2023 was less than seven, indicating homogeneity of error variances across the three experimental years.

Significant yield differences were observed in response to the f levels (p = 0.00057).

Table 3. Mean values of the parameters determined according to the depletion levels of soil available water capacity (AWC). Yield—Y; post-emergence irrigation period—PEIP; number of irrigations—NIR; applied water level in cycle—AWL; water use efficiency—WUE.

f	Y				PEIP	NIR	AWL	WUE
	2021	2022	2023	Average
(%)	(kg ha−1)				(Day)	(ud)	(mm)	(kg ha−1 mm−1)
20	6991	8470	5338	6933 a	102 a	38 a	498 a	13.9 a
40	7136	7857	5449	6814 a	100 a	21 b	486 ab	14.0 a
60	4715	7293	5065	5691 b	94 b	14 c	459 ab	12.4 a
80	4994	7274	5165	5811 b	90 b	11 c	445 b	13.1 a

Different letters in a column indicate significant differences by Tukey’s test at p < 0.05.

shows that for f = 20% and 40%, there was no significant difference in wheat grain yield, while at f = 60% and 80%, yields were significantly reduced (p < 0.05). Conversely, the post-emergence irrigation period (PEIP) significantly decreased (p = 0.0007) at f levels exceeding 40%, a trend observed concurrently with a reduction in crop yield. The NIR and AWL were also affected (p < 0.05) by the variation in f (Table 3). As f increased (p = 0.000006), NIR decreased, while AWL was highest and lowest (p = 0.024) at f = 20% and 80%, respectively. Table 3 also shows that WUE varied from 12.4 to 14.0 kg ha⁻¹ mm⁻¹, although the difference between these values was not significant in any of the f treatments (p = 0.07) (Table 3).

The experimental data to estimate Kc for wheat were obtained in 2022, in the f = 40% treatment, where wheat yield remained at the maximum level, statistically, indicating that water restriction was not relevant for crop development (Table 3). The regression of Kc values as a function of days after emergence (DAE) was optimally modeled by a quadratic polynomial model (R² = 0.75), with statistically significant coefficients (p < 0.05) (Figure 3). The Kc values were low at the beginning of the cycle, increased to a peak of 1.31 at 62 DAE, and decreased thereafter. Highest values were observed between late flowering and beginning of grain filling (56–68 DAE).

3.2. Daily N₂O and CH₄ Emissions in Response to Different Soil Water Depletion Levels and Environmental Variables

In the agricultural system under study, irrigated winter wheat was planted after rainfed soybean (growing season). In the evaluated period (23 April 2022–15 March 2024), rainfall was 2005 mm, and the mean air temperature was 21.9 °C (Figure 4a). In 2022, the water level applied to wheat ranged from 502.44 mm/cycle at 80% depletion to 587.20 mm/cycle at 20%, respectively, with a mean reference evapotranspiration (ETo) per cycle of 4.17 mm day⁻¹. In the growing season 2023, the applied water varied from 385 mm day⁻¹ at 80% depletion to 417.40 mm day⁻¹, at 20%, with ETo of 3.82 mm day⁻¹.

Daily mean N₂O and CH₄ emissions during the wheat growth periods of 2022 and 2023 were measured at 20%, 40%, 60%, and 80% water depletion. In 2022, N₂O emissions ranged from 103.80 to 129.90 µg m⁻² h⁻¹, with specific values of 120.33, 103.80, 128.0, and 129.90 µg m⁻² h⁻¹ for the respective depletion levels. CH₄ fluxes for 2022 were −4.95, −36.69, 30.20, and 0.002 µg m⁻² h⁻¹, respectively. For the 2023 wheat crop, daily mean N₂O fluxes were 73.59, 67.85, 65.77, and 71.64 µg m⁻² h⁻¹, while CH₄ fluxes were −38.05, −23.07, −18.76, and 1.62 µg m⁻² h⁻¹ for the corresponding treatments (Figure 4b,c). The mean WFPS in the 80 and 20% treatments, respectively, was 45 and 51% in 2022 and 43 and 53% in 2023 (Figure 4d).

During the period of nitrogen fertilization, in 2022 and 2023, daily fluxes were highest in the 80% treatment, with 301.4 and 150.0 µg m⁻² h⁻¹, two and four days after fertilization, with WFPS values of 47 and 38%, respectively. Additionally, the highest daily N₂O fluxes (301.4 and 236.0 µg m⁻² h⁻¹, at 80%), from wheat planted in May 2022, were associated with nitrogen fertilization and/or irrigation, respectively, observed two and seven days after fertilization, when WFPS ranged from 47 to 38%, with NO₃⁻ contents of 11.6 mg kg soil⁻¹ and NH₄⁺ of 58.0 mg kg soil⁻¹. In the growing season 2023, the highest daily N₂O fluxes (149.8 and 141.0 µg m⁻² h⁻¹, at 80%) observed in wheat were also associated with nitrogen fertilization and/or irrigation, but with WFPS of 47 to 44%, respectively, observed 4 and 17 days after fertilization, with NO_3- contents of 5.3 mg kg soil⁻¹ and NH₄⁺ of 11.1 mg kg soil⁻¹.

In growing season 2022, the mean NO₃⁻ contents were 3.4, 3.1, 3.0, and 2.7 mg kg⁻¹ (Figure 5a), and the mean NH₄⁺ contents 11.5, 9.1, 9.9, and 10.4 mg kg⁻¹ (Figure 5b), at 20, 40, 60, and 80% AWC depletion, respectively. In growing season 2023, the mean NO₃⁻ contents were 3.6, 4.5, 5.1, and 5.3 mg kg⁻¹ (Figure 4a), and the mean NH₄⁺ contents 6.6, 8.4, 9.9, and 11.0 mg kg⁻¹ (Figure 5b), at 20, 40, 60, and 80%, respectively. The soil temperature in the wheat cultivation period (2022–2023) ranged from 18.4 to 19.1 °C, with means of 18.5, 18.9, 19.0, and 19.0 °C, in the AWC depletion treatments of 20, 40, 60, and 80%, respectively (Figure 5c).

Table 4. Pearson correlation of mean N₂O and CH₄ emissions with environmental variables for year 1 (2022/2023) and year 2 (2023/2024).

	CH4	Soil Temperature	WFPS	NO3−	NH4+
2022/2023
N2O	0.12 **	0.02	0.08 *	0.09 *	0.01
CH4		−0.03	0	0.06	0.03
Soil Temperature			−0.06	0	0.12 **
WFPS				0.27 ***	0.05
NO3−					0.33 ***
2023/2024
N2O	0.16 **	0.35 ***	0.14 *	0.06	0.03
CH4		0.26 ***	0.08	0.03	0.03
Soil Temperature			0.24 ***	0.14 *	0.07
WFPS				0.12 *	0.04
NO3−					0.89 ***

Note: N₂O e CH₄ (µg m⁻² h⁻¹); soil temperature (°C); NH₄⁺ (mg kg⁻¹); NO₃⁻ (mg kg⁻¹); WFPS (%); *** significant correlation at 0.001%; ** significant correlation at 0.01% and * significant correlation at 0.05%. n = 127.

presents the correlation between N₂O and CH₄ emissions and environmental variables. Overall, the correlations were low, although some were significant. In year 1, N₂O showed a significant correlation with CH₄, WFPS, and NO₃⁻. Mineral N forms were also associated with soil temperature (NH₄⁺) and WFPS (NO₃⁻). The two years exhibited different patterns; in year 2, N₂O showed a significant correlation with CH₄, soil temperature, and EPPA. For CH₄, a significant correlation was observed with soil temperature.

3.3. Cumulative N₂O and CH₄ Emissions, Global Warming Potential (GWP), Greenhouse Gas Emission Intensity (GHGI), and Soil Enzymes

Cumulative CH₄ emissions of both growing seasons and in the mean, in response to the depletion levels, differed significantly (p = 0.0131) in growing season 2 only, at 20% depletion. Overall, cumulative CH₄ emissions ranged from −2.25 to −0.44 kg ha⁻¹ among the treatments. For N₂O, the growing seasons and means differed significantly (p = 0.0002; p = 0.013; p = 0.0001, respectively). Results showed that in growing season 1, cumulative emissions from the 60% treatment (5.53 kg ha⁻¹) were highest, followed by 80% (4.66 kg ha⁻¹), while the 20 and 40% treatments emitted 35.8 and 44.0% less, respectively, compared to the 60% treatment. In growing season 2, the 60% was equal to the 80% treatment, with higher cumulative emissions (3.56 and 3.20 kg ha⁻¹, respectively). The cumulative emissions were lower (30.90 and 23.82% less) at 20 and 40%, respectively, compared to the cumulative emissions at 60% and 80%. In the mean, cumulative emissions were highest at 60% (4.55 kg ha⁻¹) (Figure 6a–c).

The GWP for irrigated wheat differed statistically, according to the soil water depletion level, year of cultivation and mean performance (p = 0.003; p = 0.0009; p = 0.0001) (

Table 5. GWP and GHGI for irrigated winter wheat grown in 2022, in 2023, and means, at different depletion levels of available water capacity (AWC), in Planaltina, DF, Brazil.

Year	Treatment	GWP (kg CO2 eq. ha−1)	GHGI (kg CO2 eq. Mg−1)
2022	20%	878.68 (±72.74) c	104.76 (±9.73) bc
	40%	760.91 (±89.86) c	96.68 (±8.55) c
	60%	1441.88 (±98.89) a	200.73 (±37.36) a
	80%	1174.74 (±68.64) b	161.69 (±7.94) ab
2023	20%	630.09 (±28.92) b	118.04 (±4.95) b
	40%	632.11 (±40.09) b	116.53 (±13.79) b
	60%	929.68 (±70.77) a	184.59 (±13.03) a
	80%	829.31 (±66.10) a	162.80 (±19.62) ab
Mean	20%	754.39 (±49.60) c	111.40 (±2.52) b
	40%	696.51 (±58.27) c	106.61 (±10.99) b
	60%	1185.78 (±47.00) a	192.66 (±15.79) a
	80%	1002.02 (±53.13) b	162.24 (±19.62) a

Different letters in a column indicate significant differences by Tukey’s test at p < 0.05. GWP = global warming potential of N₂O and CH₄; GHGI = GHG emission intensity per yield product.

). Among the evaluated AWC depletion treatments for winter wheat in 2022, the following relationships were observed: 20% = 40%; 60% > 80, 20, and 40%; 80 ≠ 20% and 40%. The GWP was highest for the 60% treatment (1441.88 kg CO₂ eq ha⁻¹). However, in 2023, the relationships between the treatments were different: 60% = 80% ≠ 20% and 40%; 20% = 40%. The 60 and 80% treatments had the highest GWP (929.68 and 829.31 kg CO₂ eq ha⁻¹). In the mean, the GWP was higher at 60% (1185.78 kg ha⁻¹) than in the 20, 40, and 80% treatments, i.e., 36.4% higher than at 20% (Table 5).

Grain yield did not differ significantly between the years with wheat cultivation (Table 5), but GHGI differed between the evaluated years (2022 and 2023) and mean perfomance (p = 0.0036; p = 0.0084; p = 0.014, respectively). In 2022, values ranged from 96.68 to 200.73 kg CO₂ eq Mg⁻¹ grain at depletion levels of 40 and 60%, respectively. The 40% treatment was more efficient than the 60 and 80% depletion levels; i.e., the efficiency of converting gas emission per product unit was greater. For 2023, GHGI was 116.53 and 184.59 kg CO₂ eq Mg⁻¹ grain in the 40 and 60% treatments, respectively. In the mean, the 60 and 80% treatments were less efficient than 20 and 40% (higher GHGI; p = 0.014) (Table 5). The mean GHGI pattern in response to the levels of soil water depletion was clear: the 20 and 40% levels, which represent a higher frequency of irrigation and, thus, lower AWC depletion were equal, while the 60 and 80% levels were also equal, characterized by greater depletion of AWC and, consequently, less frequent irrigations and greater water replenishment.

No significant effect of the soil water depletion levels on the activity of the analyzed enzymes was observed (p = 0.90197; p = 0.79818 for arylsulfatase and β-glucosidase, respectively) (Figure 7). The mean enzymatic activities for arylsulfatase and β-glucosidase, respectively, were 49.3 and 148 µg of p-nitrophenol g⁻¹ of soil.

Table 6. Mean grain yield values, GWP, GHGI, and WUE for irrigated winter wheat grown at different depletion levels of available water capacity (AWC), in Planaltina, DF, Brazil.

Treatments (%)	Yield * (Mg ha−1)	GWP ** (kg CO2 eq. ha−1)	GHGI ** (kg CO2 eq. Mg−1)	WUE * (kg ha−1 mm−1)
20	6.93	754.39	111.40	13.9
40	6.81	696.51	106.61	14.0
60	5.69	1185.78	192.66	12.4
80	5.81	1002.02	162.24	13.1

* Data from years 2021, 2022, and 2023. ** Data from years 2022 and 2023.

summarizes the key variables evaluated. Treatments with lower soil water depletion levels (20% and 40%) achieved the highest grain yields (6.93 and 6.81 Mg ha⁻¹, respectively), the lowest total greenhouse gas emissions (GWP of 754.39 and 696.51 kg CO₂ eq ha⁻¹), and the lowest emission intensities per unit of production (GHGI of 111.40 and 106.61 kg CO₂ eq Mg⁻¹), combined with water use efficiencies of 13.9 and 14.0 kg ha⁻¹ mm⁻¹, respectively. In contrast, higher depletion levels (60% and 80%) resulted in reduced yields, increased emissions, and no significant differences in water use efficiency compared with the other treatments. Overall, irrigation management with 40% depletion of available water capacity appears to provide the optimal balance between productivity, environmental sustainability, and efficient water use.

4. Discussion

4.1. Effects of AWC Depletion on Yield

The findings indicate that soil moisture at the time of irrigation has a direct influence on wheat grain yield, possibly due to the lower water and nutrient availability, as soil matric potential increases under drier conditions. Water stress can impair photoassimilate translocation and diminish photosynthetic activity, consequently decreasing yields [41]. It has been widely reported that a reduction in available water capacity below a certain level causes a drop in wheat yield [42,43,44,45].

These results are corroborated by several studies [46], demonstrated in Kafr El-Sheikh, Egypt, that wheat yield responded differentially to different f levels and observed optimal productivity at f = 40%. Subsequent research at the same location by [47] confirmed this yield pattern in response to AWC depletion, but identified the best response at f = 50%. Contrastingly, ref. [48] demonstrated significant yield variation in winter wheat under differing f levels in semiarid Ethiopia and observed maximal yields at f = 50%. According to [30], f = 55% is recommendable for wheat irrigation management. In this study indicates that a 40% water depletion level can serve as an effective criterion for irrigation management, enabling water and energy savings without compromising maximum crop yield. This f value is therefore an efficient indicator of the most appropriate time for irrigation of winter wheat in the Cerrado of Central Brazil.

4.2. CH₄ and N₂O Emissions and Environmental Variables

The results showed that GHG emissions (N₂O and CH₄) from the Cerrado were influenced by water management, soil variables, climate, and the evaluated growing seasons. Soybean will not be discussed individually, since it was impossible to establish the AWC depletion limits during the rainy season. However, since soybean is the preferred subsequent crop to wheat, N₂O and CH₄ emissions, which influence the annual and mean emissions from the system, were taken into account for the annual metric.

The use of irrigation in agriculture can lead to either CH₄ emission or uptake (Figure 4c), as changes in soil chemistry may inhibit the activity of oxidizing enzymes [49]. Different soil water depletion levels can create conditions that favor CH₄ uptake, depending on the balance between the duration of methanotrophic oxidation and methane production, which often occur simultaneously [50].

The two growing seasons evaluated in this study were not sufficient to observe differences in CH₄ in response to the tested AWC depletion levels (Figure 5). The high drainability of the studied Oxisol [51], which does not favor anoxic conditions, should also be taken into consideration. The water that reaches the soil will restore the available water capacity to a certain level that will not favor prolonged anaerobic conditions, since the idea is to replenish the available water for plant consumption. Soil moisture, represented here by WFPS, is a critical factor for CH₄ emission/uptake, since soil diffusivity is controlled by moisture [50].

The role of the soil as a CH₄ sink, even under irrigated systems, can be attributed to its good drainage, moderate moisture levels (WFPS < 60%), and the potential for lower NH₄⁺ accumulation conditions that favor methanotrophic activity and inhibit CH₄ production through methanogenesis [52].

Different studies have mentioned the absence of a clear pattern of CH₄ uptake [53,54]. In this study and in agreement with previous research, the soil acted as CH₄ sink, in response to the application of low N rates [7,50,55]. In this study, 120 kg N ha⁻¹ was applied, which, due to the given soil conditions, induces CH₄ uptake, as occurs after low N application rates [56]. According to [7], who applied N rates of 240 kg ha⁻¹, CH₄ uptake was higher than reported by [50,55], who applied N rates of more than 240 kg ha⁻¹.

In all treatments of this study, the soil served as a CH₄ sink. This may have stimulated the activity of oxidizing microorganisms [57]. In addition, the correlation of CH₄ with soil temperature was significant in growing season 2 (p < 0.001), which suggests an influence on CH₄ production, as described by [58]. In the two studied growing seasons, N₂O was also significantly (p < 0.01) correlated (0.12 ** and 0.16 **, respectively), linking the relationship between CH₄ and N₂O that intensifies soil microbial oxidation and gas diffusion via soil pores [59]. The reason is that N₂O is an intermediate product in the nitrogen cycle, as also reported by [7].

For N₂O, maximum daily fluxes were observed after N fertilization of wheat (Figure 4b). Evidence shows that the practice of N fertilization increases N₂O emissions [6,60,61], mainly when associated with irrigation [62]. For [63], fertilization rates of more than 130 kg N ha⁻¹ raise the daily N₂O fluxes. In this study, the applied 120 kg N ha⁻¹ rate was very close to this limit, along with regular soil water replenishment. This fertilizer rate raises soil mineral N, promoting substrate for N₂O production by nitrification and denitrification processes, as well as soil biological activity [64,65]. According to [66], in well-drained soils, such as Latosolos, anaerobic sites tend to be scarce, which limits N₂O production by denitrification and suggests that nitrification predominated under the studied conditions.

The dominant nitrogen form in soil was NH₄⁺. Research [67] shows that urea application, coupled with irrigation, significantly elevates NH₄⁺ concentrations in the initial weeks post-fertilization, particularly in acidic soils where nitrification rates are retarded. In soils with low pH and reduced temperatures at the beginning of the growing season, ammonium (NH₄⁺) tends to predominate in the soil profile. Under these conditions, wheat has a good capacity to absorb and respond to nitrogen in the ammonium form. Nitrate (NO₃⁻) was probably rapidly taken up by plants, since N topdressing in wheat is applied in the tillering phase, to optimize leaf area development and grain filling. In addition, daily emissions are affected by temporal variability. Ref. [68] studied different nitrogen fertilizer rates in an African savannah and observed that more than 60% of the fluxes occurred within three weeks after fertilization. However, the fluxes were only detectable after a rain event, when soil moisture increased. Generally, the response to N from fertilization in terms of N fluxes occurs between the first and second week after application and usually disappears within two months [69].

We therefore believe that, in the mean, this depletion level can increase N₂O emissions, suggesting interactions with soil and climate variables. Thus, N₂O was significantly correlated with WFPS and NO₃⁻ in growing season 1 (2022) and soil temperature with WFPS in growing season 2 (2023) (p < 0.05), as reported elsewhere [34,58,61,70].

Elevated N₂O emissions under 60% soil water depletion may be attributed to the cyclical alternation of soil wetting and drying (Birch effect), which facilitates nitrate accumulation and enhances subsequent denitrification processes. This is corroborated by the significant positive correlations observed between N₂O emissions and both WFPS and soil temperature. Some explanations can support the higher N₂O emissions observed at 60% depletion of available soil water, such as the balance between soil moisture and oxygen availability that favors microbial nitrification and denitrification processes. At this intermediate moisture level, the soil retains sufficient oxygen for nitrification, during which N₂O can be produced as a byproduct, while transient microanoxic zones form in pores partially filled with water, enabling partial denitrification that also generates N₂O prior to its complete reduction to N₂. This balance between aerobic and partially anoxic conditions creates an environment conducive to maximal emissions of this gas, as neither full reduction to N₂ occurs nor is microbial activity limited by water or oxygen scarcity [71].

Studies conducted in different regions worldwide have demonstrated varied patterns of greenhouse gas (GHG) emissions in response to irrigation and soil management practices. In the semi-arid regions of northern China, ref. [8] reported N₂O emissions ranging from 0.85 to 1.7 kg N₂O ha⁻¹ year⁻¹ in maize crops under different irrigation and fertilization regimes, with the highest fluxes occurring under high soil moisture and elevated nitrogen application rates. These values are lower than those observed in the present study under depletion of available water capacity (AWC), likely due to differences in the irrigation method employed. Ref. [72] observed that nitrous oxide emissions in maize cultivation in southern Italy are influenced by soil texture, with higher fluxes occurring in clayey soils. The interaction between soil moisture and temperature was also relevant: under drier soil conditions, temperature played a more prominent role in intensifying emissions, whereas this effect was attenuated in wetter soils. Moreover, under high moisture conditions, sandy loam soils exhibited faster N₂O release compared to clayey soils. The authors suggest that practices such as careful irrigation management and the use of fertilizers containing nitrification inhibitors may help reduce greenhouse gas emissions in agriculture.

In general, N fertilization induces the highest N₂O emissions in the tropics [33,61]. However, it is worth recalling that in soybean–wheat systems, soybean is planted in the summer without N fertilization but contributes to N₂O emissions due to the decomposition of the crop residues, which have a low C/N ratio and nodule senescence [51]. In the planting window for irrigated winter wheat in May, thermal amplitudes are greater (10–15 °C), and 100% irrigation is applied, which leads to constant wetting, drying and rewetting (Birch effect) and, combined with soybean crop residues, provides favorable conditions for emissions [73]. In this context, the mean values for cumulative N₂O emissions were higher (2.87–4.55 kg ha⁻¹) than reported by [7] for irrigated wheat without a subsequent crop in China (0.96–1.81 kg ha⁻¹). In a soybean/rainfed maize rotation in the Brazilian Cerrado, cumulative N₂O emissions were 3.47 kg ha⁻¹ [33]. Our results with wheat under the tropical conditions of Central Brazil may have innovative applications, since the crop is becoming an alternative production option in these regions.

4.3. Global Warming Potential, Greenhouse Gas Emission Intensity (GHGI) and Soil Enzymes

The evaluation of yield, irrigation frequency, and water use efficiency (WUE) (Table 3) showed that, with increasing soil water depletion, the number of irrigations and water level applied in a cycle decreased. As an indicator of the crop response capacity to water stress, with regard to potential yield [56], WUE was higher at 40 than at 60% depletion but did not differ from 20 and 80%. At the same levels of soil water depletion, for wheat cv. BRS 394 [15], no statistical difference (p > 0.05) was found, with a lower mean value (11.8) than in this study (12.3–14.0). The reduction in irrigation applications and increase in the applied water level per operation were related to the definition criterion for irrigation, based on the level of water depletion in the soil, with the same water level as that required to restore field capacity in the explored soil profile, at the effective root depth.

In wheat cultivation, the amount of synthetic N fertilizers applied to the soil has continuously increased, resulting in N₂O and CO₂ emissions [74]. The reason is that synthetic N fertilizers tend to be highly labile and are readily available for soil nitrification and denitrification processes. Differences were possibly caused by the applied management and edaphoclimatic conditions. With a view to establish references, the GWP was calculated as the sum of the individual equivalent contribution of CO₂, N₂O and CH₄ [58]. The positive GWP values indicated that irrigated winter wheat in the Brazilian Cerrado is a source of GHG (mean emissions 713.52–1185.78 kg CO₂ eq. ha⁻¹, at 20 and 60% depletion, respectively). A positive GWP was also reported by [58], with a variation of 245.52 to 261.13 kg CO₂ eq. ha⁻¹ after applying the same N rate as in this study (120 kg ha⁻¹) and [7] with values of 125.34 to 387.81 kg CO₂ eq. ha⁻¹, both for irrigated wheat crops. On the other hand, CH₄ mitigation was greater, which resulted in lower values than found here. Thus, the contribution from N₂O emissions to GWP was greater, since CH₄ acted as a sink with negative values, which reduced GWP.

Another parameter, in addition to cumulative N₂O emissions, is the global warming potential, which can be used as an indicator of the sustainability or efficiency of a system, as greenhouse gas emission intensity (GHGI) [33,61], which relates GHG emissions with the product (grain yield or dry weight). The results showed that, in the mean, depletion levels of 20, 40, and 80% reduced the GHGI by 30 to 42%, and that the 60% depletion treatment was the least efficient (highest GHGI; Table 4). Ref. [58] found lower GHGI (14.19–32.74 kg CO₂ eq. Mg⁻¹) than reported in this study, due to lower GWP values (245–337 kg CO₂ eq. ha⁻¹). A possible strategy to improve agricultural systems would be to reduce GHG emissions by optimizing agronomic practices that contribute to mitigating GWP and increasing yields.

The reduced GHGI observed in the 40% depletion treatment indicates an optimized production system balancing water availability, crop yield, and greenhouse gas mitigation. At this level, the system efficiently converted inputs (water and nitrogen) into grain with a lower environmental footprint per unit of product. Yield remained high and statistically comparable to the 20% depletion treatment, but with reduced total water use and lower cumulative N₂O emissions, the primary driver of GWP. Therefore, the intermediate GHGI value at 40% depletion highlights the potential to reconcile agricultural productivity with environmental sustainability by minimizing input losses and climate impacts without yield penalties. This strategy aligns with low-carbon agriculture principles and climate-smart agriculture goals, promoting adaptive management to maximize resource efficiency and reduce climate forcing [2,7].

Soil enzymes play a critical role in ecosystem functioning by catalyzing key biogeochemical reactions. However, consistent with the findings of [12] in the Cerrado biome, this study observed no significant differences in soil enzyme activity across varying levels of water depletion, as water management was limited to replenishment without surplus application. Notably, β-glucosidase and arylsulfatase were identified as enzymes of particular interest, corroborating the results reported by [75]. In Brazil, this technology has been used since 2020 for on-farm evaluations [76], mainly due to its sensitivity [77], low seasonal variability [78], ease of application, low cost [76], and correlation with several edaphoclimatic factors.

The apparent stability in enzyme activity suggests that the depletion levels applied did not significantly affect soil microbial functioning. This may be related to the absence of extreme moisture conditions, as irrigation was managed to replenish water up to field capacity without excess. Other contributing factors may include the maintenance of crop residue cover and the use of no-till systems, which promote microenvironmental stability and provide physical protection to the soil microbiota. Additionally, enzyme activities reflect processes related to the carbon and sulfur cycles, which may respond more slowly to short-term changes in soil moisture [10,11].

The findings of this study carry significant implications for sustainable irrigation management in the Cerrado biome, demonstrating that a 40% depletion of available soil water optimizes the balance between crop productivity and greenhouse gas emission mitigation, particularly N₂O. This approach results in reduced emission intensity per unit of product (GHGI) without compromising yield, establishing it as a viable technical option for low-carbon agriculture. Moreover, the absence of significant effects on soil enzyme activity indicates microbial resilience to moderate fluctuations in soil moisture, particularly under no-till management with residue retention, thereby reinforcing the functional stability of soil within tropical conservation agriculture systems.

The methodologies employed in this study can be expanded to other crops and production systems in the Cerrado, contributing to the development of functional soil indicators and the valuation of ecosystem services associated with the sustainable intensification of tropical agriculture. Despite these advances, to date no studies have been published in Brazil that explore this approach in crop sequence systems involving irrigated wheat, particularly regarding N₂O and CH₄ emissions. Thus, this study is noteworthy for addressing this research gap by evaluating, in an unprecedented manner, the application of this integrated framework in irrigated wheat systems, a strategically important crop for the Cerrado and increasingly relevant in the context of sustainable agricultural intensification.

The principles of management based on partial depletion of available soil water have scalability potential for similar agricultural systems in different regions worldwide. However, the application of these results on a global scale must consider local particularities such as soil type, climate, and production system. The scalability of efficient irrigation management requires the development of adaptive models that can incorporate these local variables to optimize the balance between productivity and environmental sustainability.

A central aspect of this study was the trade-off observed between yield optimization and GHG mitigation. Although the irrigation strategy with 40% depletion of available soil water showed the best combination of high productivity and lower relative emissions, higher irrigation levels could result in increased yields but also be accompanied by elevated N₂O emissions and negative impacts on soil biological quality, likely compromising both environmental and economic sustainability. On the other hand, more severe water restrictions compromised crop performance. Therefore, irrigation management must seek a balance that addresses both food security and the reduction of environmental impact, an essential aspect for the sustainability of agricultural systems in a global climate change scenario.

Overall, the results confirm that irrigation management differentially influences each variable analyzed, and that the optimal response is determined not solely by yield maximization but also by the reduction of emissions and the maintenance of soil biological quality. A 40% depletion of available water capacity emerged as the balance point, combining high productivity with lower environmental impact, while the hypothesis of increased emissions under lower depletion levels was not confirmed. This finding highlights the importance of understanding the interactions among water management, microbiological processes, and edaphoclimatic conditions. Such an integrated perspective reinforces the role of irrigation strategies adapted to tropical conditions as a tool for enhancing the sustainability and resilience of agricultural systems in the Cerrado.

5. Conclusions

Addressing the study objective, this research evaluated the effects of different soil water depletion levels on grain yield, greenhouse gas (GHG) emissions, water use efficiency (WUE), and soil enzymatic activity in winter wheat cultivated under no-tillage in the Brazilian Cerrado. The results indicate that the strategy based on 40% depletion of the available soil water was the most efficient for promoting sustainable production. This practice ensured high productivity and water use efficiency while significantly reducing N₂O emissions, global warming potential (GWP), and GHG intensity per ton of grain (GHGI). The soil acted as a methane (CH₄) sink and a N₂O source, with the highest emissions occurring under 60% depletion, possibly due to suboptimal moisture conditions favoring nitrification and denitrification processes. The lack of a significant response in enzyme activities suggests that short-term changes in the water regime do not strongly affect functional microbial activity in no-tillage systems. Therefore, 40% depletion represents the best balance between productive performance and environmental sustainability, fostering climate-smart agricultural practices and efficient resource use in tropical environments. Future research should assess the long-term impacts on soil health and test its applicability across crops and regions to guide incentive programs and public policies for climate change adaptation in agriculture.