Carbon Budget of Paddy Fields after Implementing Water-Saving Irrigation in Northeast China

Abstract

Water-saving irrigation is recognized as an effective agricultural management due to water security and environmental protection problems. In Northeast China, an increasing number of paddy fields are shifting from conventional irrigation to water-saving irrigation. However, there is limited knowledge regarding the carbon (C) budget of paddy fields after implementing water-saving irrigation in Northeast China. A 2-year consecutive field study was performed from 2018 to 2019 using three different irrigation regimes (conventional irrigation (FI), controlled irrigation (CI), and intermittent irrigation (II)) and two nitrogen (N) fertilization levels (110 and 165 kg N ha⁻¹) in a paddy field of Northeast China. The present study aimed to quantify the net ecosystem C budget (NECB) and net global warming potential (net GWP) after the implementation of water-saving irrigation in Northeast China. Both CI and II enhanced the C sequestration capacity of this paddy field. The net primary productivity (NPP) under CI and II was higher than FI by 18–38% and 11–33%, respectively, when the same N fertilization level was applied. The NECB ranged from 1151 to 2663 kg C ha⁻¹, indicating that all treatments acted as net C sinks. II increased the NECB through increasing NPP, which exceeded increased removal of harvest and C mineralized losses. Under II, the NECB was significantly higher than FI and CI when the same N fertilization level was applied (p < 0.05). The net GWP under II and CI was significantly lower than FI (p < 0.05). The net GWP under II was lower than CI when the N fertilization level was 165 kg N ha⁻¹, but no significant differences were detected. These results demonstrated that the II with 165 kg N ha⁻¹ of paddy fields strongly decreased net GWP in Northeast China to combat global climate change.

1. Introduction

Rice is a main food for >50% of the world’s population. As the largest rice-producing country, China shares ~19% of the rice-growing area and 32% of the rice yield in the world [1]. Given the decline in the availability of water resources and increasing demands and environmental benefits for water-saving irrigation, several paddy fields with conventional irrigation (FI) have been replaced by water-saving irrigation in China [2,3]. Some widely used water-saving irrigation regimes, including rain-gathering irrigation (RGI), controlled irrigation (CI), and intermittent irrigation (II), have been applied across China, of which CI and II are the main regimes in Northern China [4,5,6]. Previous studies demonstrated that water-saving irrigation regimes enhance C sequestration by affecting plant growth due to differences in the amount of irrigation, irrigation frequency, and irrigation intervals during the rice cropping season [7,8]. Akter et al. reported that appropriate water-saving irrigation regimes promoted root activity, which thereby benefited other physiological processes and resulted in rice growth [9]. Wang et al. found that the application of water-saving irrigation reduced N loss, and more N was absorbed for rice growth, which thereby increased crop yields [10]. Moreover, some studies have reported that water-saving irrigation will affect plant growth and greenhouse gas (GHG) emissions after its implementation [11,12,13].

The effects of irrigation on GHG emissions are a result of soil moisture regulation. It is well-known that soil moisture is a key factor in GHG emission regulation as it affects microbe activities and other relevant processes [14,15]. Water-saving irrigation improves soil aeration by regulating soil moisture. Soil aerobic conditions strengthen microbial activities and soil respiration and thus promote CO₂ production, which is emitted from the soil through the roots and soil respiration [16]. Qi et al. reported that the irrigation mode, irrigation volume, and irrigation frequency influence CO₂ emissions [17]. Yang et al. reported that CI strongly increased soil respiration, which resulted in a decrease in NEE in paddy fields [18]. Nevertheless, soil aerobic conditions significantly inhibit CH₄ emissions as they impede methanogen activities and all related processes and increase the redox potential [19]. The results of two experiments indicated that CI decreased CH₄ emissions by 34% and 83.5%, respectively, compared to FI in Northeast and Southeast China [20,21]. Previous studies investigated GHG emissions and C sequestration under different water-saving irrigation regimes. However, the C budget of paddy field ecosystems remains uncertain after water-saving irrigation is implemented. Paddy field ecosystems may act as either a C sink or C source after water-saving irrigation is implemented. Therefore, it is important to understand the balance between C sequestration and GHG emissions in paddy fields, as well as enhance the C sequestration capacity to establish water-saving agricultural production.

Aiming to meet the needs of rice yields in China, rice-growing areas have increased rapidly in Northeast China, especially in Heilongjiang Province [22]. The National Bureau of Statistics of China reported that the total paddy field area was 2.68 × 10⁶ ha in 2000 and increased at a rate of 0.16 × 10⁶ ha year⁻¹. In 2017, the total paddy field area of Northeast China reached 5.62 × 10⁶ ha. Currently, FI is used in most areas of Northeast China. Such a rapid increase in paddy field areas requires a great deal of water for agricultural irrigation. Moreover, the population growth, urbanization, and industrialization of Northeast China, along with water resource quality decline, have aggravated the dire shortage of agricultural irrigation water. Due to the severe shortage of agricultural irrigation water, government departments have attempted to promote water-saving irrigation through training and other means every year. In 2018, the rice water-saving irrigation area of Heilongjiang Province reached 1.24 × 10⁶ ha, and the annual water savings was 1.85 × 10⁸ m³ [23]. Additionally, the effects of environmental factors and basic soil physical-chemical properties on rice growth and C mineralized losses vary across different regions. Therefore, selecting the most appropriate irrigation regime is vitally important for Northeast China.

This study aimed to address the following questions: (1) Which irrigation regimes are best for Northeast China? (2) How much of an effect does water-saving irrigation have on the balance between C mineralized losses and C sequestration? We aimed to quantify the net ecosystem C budget (NECB) and the emissions of CH₄, as well as their contributions to the net GWP after implementing water-saving irrigation. The aims of this study were to (1) explore whether paddy fields act as a C sink or C source after implementing water-saving irrigation, and (2) identify which water-saving irrigation regime can decrease net GWP in Northeast China to combat global climate change.

2. Materials and Methods

2.1. Site Description

The field site was located in Suihua, Heilongjiang Province, China (46°57′28″ N, 127°40′45″ E; 168 m.a.s.l.), which is a typical black soil paddy field growing area at the intersection of the Songnen Plain and Xiao Xing’an Mountains. The site has a monsoon climate in a cold temperate zone. In this region, the average annual precipitation was 577 mm, and the average air temperature was 1.69 °C. The soil texture of this region was classified as sandy clay, and the physical-chemical properties of the 0–20 cm layer were pH 6.6 (H₂O, 1:1), 40.6 g kg⁻¹ organic matter, 186.8 mg kg⁻¹ alkaline N, 34.9 mg kg⁻¹ available p, and 106.8 mg kg⁻¹ available K.

The dominant cropping system in the region is single cropping rice, which is cultivated ~120 days during the summer and autumn. Thereafter, the soil is frozen under extremely cold conditions during the fallow season. Prior to experimentation, the test field had rice planted for >20 years and received frequent farmyard manure input. Measurements were carried out in a 2-year field experiment from 2018 to 2019. Straw was shredded (<5 cm) by an automated pulverizer and returned to the soil at a depth of 20 cm at the time of harvest.

2.2. Experimental Design

The experiment was laid out in a spilt-plot design with a water-saving irrigation regime, and N fertilization was used as sub-plots. The irrigation regimes included (1) FI, (2) CI, and (3) II (

Table 1. Different water management regime at rice growth stages.

Irrigation Regime	Turning -Green Stage	Early Tillering Stage	Middle Tillering Stage	Late Tillering Stage	Jointing -Booting Stage	Heading-Flowering Stage	Milk Stage	Yellow -Ripe Stage
FI	0~30 mm	0~50 mm	0~50 mm	Drainage	0~50 mm	0~50 mm	0~50 mm	Naturally drying
CI	0~30 mm	0.7 θs~0	0.7 θs~0	Drainage	0.8 θs~0	0.8 θs~0	0.7 θs~0	Naturally drying
II	0~30 mm	0~40 mm	0~40 mm	Drainage	0~30 mm	0~40 mm	0~40 mm	Naturally drying

Note: θs refers to soil saturated water content mass fraction in the root layer. FI, conventional irrigation; CI, controlled irrigation; II, intermittent irrigation.

). II sets the irrigation amount and intervals during the rice cropping season to make the paddy field in shallow water or no surface water conditions. CI keeps the paddy field in shallow water only in the turning green stage while allowing the paddy field under no surface water conditions throughout the rest of the cropping season. In the experimental area, the amount of N, P₂O₅, and K₂O is recommended as 110, 45, and 80 kg ha⁻¹, respectively [24,25,26,27]. The N fertilization levels consisted of (1) 110 kg N ha⁻¹ (N110) and (2) 165 kg N ha⁻¹ (N165) (>50% conventional N fertilization application; more N fertilizer is applied after returning straw) [28,29]. One day before rice transplanting, 45% of the N fertilizer was applied as basal fertilizer. Twenty-four and seventy-two days after transplanting, 20% and 35% of N fertilizer were applied as tillering and grain fertilizer, respectively. The application rates of P₂O₅ and K₂O were the same in all treatments (45 kg ha⁻¹ P₂O₅ and 80 kg ha⁻¹ K₂O). Before rice transplanting, 50% of the K₂O was applied, and the other 50% of the K₂O was applied at the leaf age of 8.5. Each treatment was laid out with 3 replicates. In total, there were 6 main-plots and 18 sub-plots. The size of the sub-plots was 10 × 10 m. Between each plot, a concrete barrier (height: 40 cm) was laid down as a barrier to prevent the exchange of water and fertilizer. Twenty-one-day-old rice seedlings (3 plants per hill, Longqing 3, China) were transplanted by automatic machine. The planting density was 16.67 cm × 30 cm. All other agricultural management practices, including raising seeding, pesticide application, and harvesting, were the same for each treatment [30]. In 2018 and 2019, the seedlings were transplanted on May 18 and May 19 and rice was harvested on September 22 and September 21.

2.3. C Mineralized Losses

Heterotrophic respiration (Rh) was measured in situ using LI-COR closed-chamber soil respiration system (LI-COR 8100; Li-cor Inc; Lincoln, MI, USA) [31]. PVC collars (10 cm inner diameter, 50 cm in height) were laid out with 3 replicates in each plot during the rice cropping season. In each collar, rice was not cultivated, and weeds were entirely removed. PVC collars were drilled to 5–50 cm, through which the management of water and fertilizer inside and outside the PVC collars were consistent, and roots were prevented from entering. PVC collars were inserted into the soil to 45 cm after rice was transplanted. Measurements were carried out from transplantation to harvest once per week. The measuring time was fixed between 9:00 and 11:00 a.m., as the Rh and CH₄ fluxes during this time quantum were close to the mean daily Rh and CH₄ fluxes [32]. In the event of heavy rainfall, the measurement time was delayed. Methane fluxes were measured at the same time on the same day using the static chamber-gas chromatography method. Static chambers were also laid out with 3 replicates in each plot during the rice cropping season. The static chambers consisted of a chamber and stainless-steel base. The chambers were made of transparent organic glass. Each chamber was equipped with an air thermometer and electronic fan. Chambers were removed from the base, except during gas collection. The stainless-steel base was embedded in each plot before transplanting with a sealing groove (3 cm width, 5 cm height) reserved at the top. While taking measurements, water was injected to seal the groove to avoid gas exchange in the chamber with outside air. At 0, 10, 20, and 30 min after chamber closure, 4 gas samples were collected using a 50 mL E-Switch gas bag via a rubber tube for each CH₄ flux measurement. The temperature inside the chamber was also recorded [33]. All gas samples were analyzed within 24 h using gas chromatography (GC-2010 Plus; Shimadzu Corporation; Kyoto; Japan). The gas chromatograph was equipped with a flame ionization detector (FID); methane concentration was analyzed at 200 °C; the carrier gas was N₂ with a purity of 99.99%.

Methane fluxes were calculated based on changes in their concentrations throughout the sampling period and estimated as the slope of the curve of the concentration versus time [34]. Then, CH₄ fluxes were calculated as follows [35]:

$$F=\rho \times h\times dc/dt\times 273/273+{\rm T}$$

(1)

where F is the CH₄ fluxes (mg m⁻² h⁻¹), ρ is the density of CH₄ under a standardized state (0.714 mg cm⁻³), h is the effective height of the chamber above the soil or surface water (m), dc/dt is the rate of increase of CH₄ gas concentrations in the chamber (g m⁻³ d⁻¹), and T is the mean air temperature inside the chamber at the time of sampling.

The seasonal Rh (kg C ha⁻¹) and CH₄ fluxes (kg C ha⁻¹) for the entire cropping period were calculated as follows [36]:

$$Seasonal Rh and C{H}_4={\displaystyle {\sum }_i^n(E_i\times D_i)}$$

(2)

where E_i is the Rh (μmol m⁻² s⁻¹) and CH₄ fluxes (mg m⁻² h⁻¹) at the ith interval during sampling, D_i is the number of days between ith sampling and (i − 1)th sampling, and n is the sampling number.

2.4. NECB Calculation

Based on a mass C budget approach, the NECB (kg C ha⁻¹) was calculated between the rate of C gain or loss in a cropping system. Previous studies demonstrated that the NECB with a positive value indicates that paddy fields act as a C sink, and a negative value indicates a C source. The NECB was calculated as follows [37,38].

$$Re=Ra+Rh$$

(3)

$$\begin{array}{c}NECB={\displaystyle \sum Cinput-{\displaystyle \sum Coutput-(NPP+Ra)-(Re+Harvest+C{H}_4)}}\\ -NPP-Rh-Harvest-C{H}_4\end{array}$$

(4)

where NPP (kg C ha⁻¹) is the net primary production during rice cropping season and the fallow season [39], and Re is the sum of plant respiration Ra and Rh [40]; Rh was measured using the above closed-chamber soil respiration system, and CH₄ fluxes were measured by the static chamber-gas chromatography method. In the event that all residues were returned to the field, only grain was moved from the field at the time of harvest. Soil C losses through runoff and leaching were typically excluded from the calculations and therefore not included in the current study.

Grain, stem, and leaf were harvested using sickle in each plot, and the size of the sampling frame was 100 × 100 cm. Roots were acquired from the 30 cm soil layer using a root extractor (5.8 cm inner diameter) in each plot and were washed using a pressure-water gun [41]. Grain, stem, leaf, and root were placed in a drying oven at 105 °C for 30 min, dried at 70 °C to a constant mass, and weighed. Grain, stem, leaf, and root were ground into a powder. Then, sample C was measured using a vario TOC elemental analyzer (Elementar vario TOC; Elementar; Hanau; Germany). The NPP of rice during rice cropping was calculated as follows [42,43,44,45]:

$$NPP=NP{P}_{grain}+NP{P}_{stem}+NP{P}_{leaf}+NP{P}_{root}+NP{P}_{litter}+NP{P}_{rhizodeposit}$$

(5)

in which rice NPP_{(grain, stem, leaf, root)} was calculated using the dried matter weight multiplied by the C content at the time of harvest. NPP_litter and NPP_rhizodepoait were estimated by other rules. Litter accounted for ~5% of the total dried matter weight, and the C leaf content was used to calculate NPP_litter [46]. NPP_rhizodepoait mainly included exudates, exfoliated root hairs, and dead root and was calculated as 11% of the total C content [47].

2.5. Net Global Warming Potential

The net GWP (kg C ha⁻¹) was used to estimate the climatic impact of the paddy field system after implementing water-saving irrigation. N₂O emissions were not considered in this study. The net GWP was calculated as follows [48]:

$$Net GWP=28\times Seasonal C H_4-44/12\times NECB$$

(6)

where the warming potential of CH₄ was 28 on a 100-year time horizon [49], Seasonal CH₄ represents cumulative fluxes of CH₄ during the entire cropping period, and NECB represents the rate of C gain or C loss in a cropping system; 44/12 was the conversion coefficient of C to CO₂.

2.6. Statistical Analysis

All formula calculations were calculated using Excel 2010. Statistical analyses were performed by SPSS v19.0 (IBM). Univariate multivariate analysis of variance was used to test the effects of irrigation regime, N fertilization, year, and their interactions on NPP and the components of NECB and net GWP. Three-way ANOVAs were used to evaluate significant differences in all treatments. The significant threshold of all statistical analyses was p < 0.05. All figures were visualized using ORIGIN v9.0 (OriginLab).

3. Results

3.1. Environmental Conditions

The typical climatic conditions of the study site were observed during the 2-year field study; environmental conditions did not differ considerably from 2018 to 2019 (Figure 1). During rice cultivation, the daily air temperature ranged from 2.0 °C to 37.6 °C in 2018 and from 1.7 °C to 33.7 °C in 2019. The seasonal precipitation in 2018 (563.6 mm) was higher during rice cultivation than in 2019 (542.5 mm).

3.2. Net Primary Productivity

The rice dried matter weight and their C contents are presented in Figure 2. Compared with FI, CI and II increased dried matter weight and rice C contents when the same N fertilization application level was applied. The increased intensity of rice dried matter weight of CI was strongest, and CI and II were higher than FI by 15–32% and 9–30%, respectively. The C contents of each part of the rice plant did not significantly differ among the three irrigation regimes. The dried matter weight and C contents of rice under N165 were higher than N110.

The NPP components are summarized in

Table 2. Changes in NPP components in paddy fields after the implementation of water-saving irrigation regime.

Year		2018										2019
Treatment		FN110	FN165		CN110		CN165	IN110		IN165		FN110	FN165		CN110		CN165	IN110		IN165
NPPgrain (kg C ha−1)		2824 ± 71e	3578 ± 126cd		3766 ± 89c		4765 ± 119a	3490 ± 106d		4290 ± 87b		1928 ± 42g	2538 ± 53f		2985 ± 98e		3580 ± 186cd	2829 ± 97e		3573 ± 114cd
NPPstem (kg C ha−1)		1529 ± 38e	1934 ± 59b		1738 ± 49cd		2062 ± 61a	1685 ± 56cd		2031 ± 46ab		1285 ± 34f	1693 ± 58cd		1687 ± 53cd		2046 ± 41a	1648 ± 29d		1781 ± 46c
NPPleaf (kg C ha−1)		514 ± 13g	664 ± 14cd		644 ± 20de		741 ± 19a	621 ± 21e		684 ± 16bc		402 ± 12h	613 ± 16e		562 ± 17f		702 ± 19b	522 ± 15g		632 ± 17de
NPProot (kg C ha−1)		610 ± 13b	716 ± 22a		465 ± 17e		569 ± 21c	507 ± 15d		627 ± 17b		472 ± 17de	627 ± 15b		416 ± 17f		507 ± 16d	452 ± 14e		597 ± 18bc
NPPlitter (kg C ha−1)		268 ± 6e	339 ± 10c		327 ± 8cd		402 ± 10a	314 ± 10d		371 ± 9b		205 ± 5f	270 ± 7e		279 ± 9e		339 ± 9c	272 ± 7e		321 ± 9cd
NPPrhizodeposit (kg C ha−1)		602 ± 14e	758 ± 24c		727 ± 17cd		895 ± 22a	693 ± 22d		839 ± 18b		450 ± 11f	602 ± 15e		622 ± 20e		752 ± 27c	600 ± 16e		724 ± 20cd
NPP (kg C ha−1)		6347 ± 148e	7988 ± 249c		7666 ± 181cd		9435 ± 234a	7310 ± 229d		8841 ± 190b		4742 ± 119f	6343 ± 159e		6552 ± 213e		7926 ± 285c	6323 ± 173e		7628 ± 211cd
Statistical analysis
Year (A)	Irrigation regime (B)			N fertilization (C)		NPPgrain (kg C ha−1)			NPPstem (kg C ha−1)		NPPleaf (kg C ha−1)			NPProot (kg C ha−1)		NPPlitter (kg C ha−1)			NPPrhizodeposit (kg C ha−1)
A (year)						***			***		***			***		***			***
B (irrigation regime)						***			***		***			***		***			***
C (N fertilization)						***			***		***			***		***			***
A × B						*			**		ns			**		ns			ns
A × C						*			ns		**			ns		ns			ns
B × C						ns			**		***			ns		ns			ns
A × B × C						ns			*		ns			ns		ns			ns

Different lowercase letters indicate significant differences between treatments from 2018 to 2019 (p < 0.05). Values are presented as the mean ± standard error of 3 replicates. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, non-significant F-values. NPP, net primary productivity; NPP_grain, NPP_stem, NPP_leaf, NPP_root, NPP_litter, and NPP_rhizodeposit, net primary productivity of grain, stem, leaf, root, litter, and rhizodeposit.

. NPP_grain, NPP_stem, NPP_leaf, NPP_litter, NPP_rhizodeposit, and NPP under CI and II were significantly higher than FI when the same N fertilization level was applied (p < 0.05). Additionally, NPP_grain, NPP_stem, NPP_leaf, NPP_litter, NPP_rhizodeposit, and NPP under CI were higher than II. NPP under CI and II were higher than FI by 18–38% and 11–33%, respectively. However, NPP_root under FI was higher than CI and II. NPP_grain, NPP_stem, NPP_leaf, NPP_root, NPP_litter, NPP_rhizodeposit, and NPP under N165 were all significantly higher than N110. These changes exhibited similar trends in 2018 and 2019. Nonetheless, NPP_grain, NPP_stem, NPP_leaf, NPP_litter, NPP_rhizodeposit, and NPP of each treatment in 2018 were higher than in 2019.

3.3. Carbon Mineralized Losses

Carbon mineralized loss mainly consisted of Rh and CH₄ fluxes and was visualized during the rice cropping season (Figure 3). The CH₄ fluxes of each treatment sharply increased after the turning green stage, where the highest CH₄ fluxes were observed at the late tillering stage. Thereafter, CH₄ fluxes dramatically decreased at the drainage stage and reached another peak at the jointing-booting stage. Except for the drainage period, CH₄ fluxes under FI were higher than CI and II when the same N fertilization level was applied. Methane fluxes reached up to 55.53 mg m⁻² h⁻¹ in FN165. Rh in each treatment increased after the turning green stage, where the highest Rh fluxes were observed at the late tillering stage and decreased to a low level at the drainage stage. During rice cropping, Rh under CI and II was higher than FI when the same N fertilization level was applied. Rh reached 9.13 μmol m⁻² s⁻¹ in CN165. Furthermore, the Rh and CH₄ fluxes under N165 were higher than N110 under all irrigation regimes.

3.4. Net Ecosystem Carbon Budget

Seasonal Rh, CH₄ fluxes, and NECB are summarized in

Table 3. Seasonal Rh, CH₄ fluxes, NECB, and net GWP in a paddy field after the implementation of water-saving irrigation.

Year	Treatment	C Input (kg C ha−1)		C Output (kg C ha−1)				NECB (kg C ha−1)	net GWP (kg C ha−1)
		NPP	Sum	Harvest	Rh	CH4	Sum
2018	FN110	6437 ± 148e	6437 ± 148e	2824 ± 71e	758 ± 155ef	869 ± 20b	4451 ± 244d	1896 ± 97c	17371 ± 252b
	FN165	7988 ± 249c	7988 ± 249c	3578 ± 126cd	1000 ± 150def	960 ± 28a	5538 ± 291c	2450 ± 76b	17915 ± 543ab
	CN110	7666 ± 181cd	7666 ± 181cd	3766 ± 89c	2338 ± 166b	389 ± 19e	6493 ± 188b	1173 ± 8f	4769 ± 528de
	CN165	9435 ± 234a	9435 ± 234a	4765 ± 119a	2709 ± 145a	482 ± 21d	7956 ± 281a	1478 ± 123d	8067 ± 749cd
	IN110	7310 ± 229d	7310 ± 229d	3490 ± 106d	1067 ± 142de	516 ± 17cd	5072 ± 247c	2238 ± 73b	6242 ± 246e
	IN165	8841 ± 190b	8841 ± 190b	4290 ± 87b	1324 ± 143d	564 ± 19c	6178 ± 249b	2663 ± 64a	6037 ± 416e
2019	FN110	4742 ± 119f	4742 ± 119f	1928 ± 42g	703 ± 140f	856 ± 40b	3488 ± 222e	1255 ± 115ef	19377 ± 819a
	FN165	6343 ± 159e	6343 ± 159e	2538 ± 53f	921 ± 164ef	885 ± 39b	4344 ± 256d	1999 ± 112c	17441 ± 862b
	CN110	6552 ± 213e	6552 ± 213e	2985 ± 98e	1931 ± 156c	484 ± 37d	5401 ± 271c	1151 ± 134f	9341 ± 1133c
	CN165	7926 ± 285c	7926 ± 285c	3580 ± 186cd	2430 ± 156ab	513 ± 37cd	6521 ± 378b	1404 ± 96de	9196 ± 962c
	IN110	6323 ± 173e	6323 ± 173e	2829 ± 97e	1008 ± 162def	513 ± 36cd	4351 ± 282d	1973 ± 122c	7131 ± 801de
	IN165	7628 ± 211cd	7628 ± 211cd	3573 ± 114cd	1277 ± 153d	535 ± 37cd	5386 ± 275c	2243 ± 113b	6757 ± 998de
Statistical analysis
A (year)		***	***	***	*	ns	***	***	**
B (irrigation regime)		***	***	***	***	***	***	***	***
C (N fertilization)		***	***	***	***	***	***	***	ns
A × B		ns	ns	*	ns	**	ns	***	ns
A × C		ns	ns	*	ns	ns	ns	ns	ns
B × C		ns	ns	ns	ns	ns	ns	**	ns
A × B × C		ns	ns	ns	ns	ns	ns	ns	ns

Different lowercase letters indicate significant differences between treatments from 2018 to 2019 (p < 0.05). Values are presented as the mean ± standard error of 3 replicates. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, non-significant F-values. Rh, heterotrophic respiration; CH4, seasonal CH4 fluxes; NPP, net primary productivity; NECB, net ecosystem carbon budget; net GWP, net global warming potential.

. Seasonal Rh ranged from 703 to 2709 kg C ha⁻¹ at the rice cropping stage. Seasonal Rh of CN165 was the highest in 2018, while FN110 was the lowest in 2019. Seasonal Rh under CI was significantly higher than FI and II when the same N fertilization level was applied (p < 0.05). Seasonal Rh under CI was 2.64–3.08 and 1.90–2.19-times higher than FI and II, respectively. Seasonal CH₄ fluxes ranged from 389 to 960 kg C ha⁻¹ at the rice cropping stage. Seasonal CH₄ fluxes of FN165 were the highest, while CN110 was the lowest in 2018. Seasonal CH₄ fluxes under FI were significantly higher than CI and II when the same N fertilization level was applied (p < 0.05). Seasonal CH₄ fluxes under FI were 1.65–2.23 and 1.65–1.70-times higher than CI and II, respectively. Moreover, seasonal Rh and CH₄ fluxes under N165 were higher than N110 under all irrigation regimes.

The NECB ranged from 1151 to 2663 kg C ha⁻¹, indicating that all treatments acted as a net C sink. The NECB of IN165 was the highest in 2018, while CN110 was the lowest in 2019. Results revealed that NECB under II was significantly higher than FI and CI when the same N fertilization level was applied (p < 0.05). The NECB under II was higher than FI and CI by 9–57% and 60–91%, respectively. The NECB under FI was higher than CI when the same N fertilization level was applied. The NECB under N165 was significantly higher than N110 under all irrigation regimes (p < 0.05).

3.5. Net Global Warming Potential

The net GWP is summarized in Table 3. The net GWP of the NECB and CH₄ fluxes ranged from 6037 to 19,377 kg C ha⁻¹. The net GWP of FN110 was the highest in 2019, while IN165 was the lowest in 2018. When the same N fertilization level was applied, net GWP under II and CI were significantly lower than FI (p < 0.05). The net GWP under II was also lower than CI when the N fertilization level was 165 kg N ha⁻¹, but no significant differences were detected. Moreover, the net GWP under N165 was lower than N110 under II.

4. Discussion

Whether paddy field ecosystems function as C sinks or sources is related to the balance of organic C inputs and outputs. It is well-known that appropriate agricultural practice can enhance the ability of C sequestration and mitigate CO₂ emissions to combat climate change [50]. In Northeast China, due to the continuous expansion of paddy rice areas, it is imperative to implement water-saving irrigation. Government departments reported that the rice water-saving irrigation area of Heilongjiang Province will exceed 2.8 × 10⁷ hm² by 2025, accounting for ~70% of the rice-growing area [51]. However, the status of the C budget in paddy fields remains uncertain. In paddy field ecosystems, Re and C sequestration in rice plants are likely the main processes affecting the C balance. However, additional C inputs (only contain organic fertilizer) and outputs (grain, even straw, except for residues) also affect the C balance of paddy field ecosystems [52,53]. The calculation of NECB accounts for all C inputs and outputs to evaluate the C balances of paddy field ecosystems. The NECB has been recognized as a scientific evaluation index for determining short-term C balances [54,55].

In single cropping rice paddy ecosystems, more organic C inputs can increase rice dried matter weight, which indicates that SOC stock increases [56,57]. In this study, no additional C inputs were detected, and the C inputs only contained NPP. Results revealed that NPP under CI and II was higher than FI by 18%–38% and 11%–33%, respectively. This may be because CI and II increased the rice dried matter weight and C contents. These results confirmed that an increase in rice dried matter weight and C contents resulted in increased C inputs. Kim et al. reported that the higher-level N fertilization significantly increased NPP, mainly through the increase of rice dried matter weight [39]. These findings were consistent with the results of this study. This study showed that NPP under N165 was higher than N110.

In single cropping rice paddy ecosystems, harvest removal (grain, including straw, except for residues) and C mineralized loss are the main C output sources [58,59]. In China, the policy stipulates that anyone who burns straw will be punished, so most farmers return all straw to the field at the time of harvest. In this region, grain was removed at the time of harvest, and harvest removal only contained NPP_grain. Approximately 55–69% of C outputs were covered by harvest removal, while 30–45% were covered by C mineralized loss. These results indicated that C mineralized loss accounts for a small proportion of C outputs relative to harvest removal; thus, the contribution of C mineralized loss to NECB was smaller than NPP_grain.

Rh and CH₄ are the main pathways of C mineralized loss from soil to the atmosphere in paddy field ecosystems [60]. Seasonal Rh under CI was 2.64–3.08 and 1.90–2.19-times higher than FI and II, respectively. No surface water management of water-saving irrigation resulted in a higher Rh in the paddy field. The soil oxygen content increases due to frequent alternate wet and dry conditions. The soil organic matter decomposition was promoted mainly through the increases of the soil oxygen content and the redox potential, and its products provide more substrates for Rh. Finally, water-saving irrigation keeps the paddy field in shallow water or no surface water conditions and reduces the distance of CO₂ emissions into the atmosphere [18]. In this study, seasonal CH₄ fluxes under FI were 1.65–2.23 and 1.65–1.70-times higher than CI and II, respectively. Methane production is mainly derived from anaerobic methane bacteria. FI provides a long-term anaerobic environment that promotes methane production [61]. However, water-saving irrigation improves soil aeration, which greatly reduces CH₄ production.

The year, irrigation regime, and N fertilization application level of each factor had a significant effect on NECB, as well as the interaction between year and N fertilization application level, and the interaction between irrigation regime and N fertilization application level (Table 3). The NECB under N165 was significantly higher than N110 under all irrigation regimes. This indicates that the N fertilization level can promote the C sink capacity of paddy fields. The NECB under II was higher than FI by 9–57%. These results indicated that the NECB increased after the implementation of II, mainly because the increased NPP outweighed the increased harvest removal. However, the NECB under FI was higher than CI. This may be because the NPP_grain of CI was significantly higher than FI.

The year and irrigation regime of each factor had a significant effect on net GWP (Table 3). The net GWP decreased significantly after the implementation of water-saving irrigation due to increased C losses, which were far lower than the decreased CH₄ fluxes. These results confirmed our hypothesis that water-saving irrigation could decrease net GWP. Wu et al. found that NECB with a positive value can contribute to mitigating GHG emissions [62]. This is consistent with our results, which showed that net GWP under II was lower than FI due to the higher NECB under II. However, the NECB under CI was lower than FI when the same N fertilization level was applied, but its net GWP was lower than FI. This may be because seasonal CH₄ fluxes under CI were significantly lower than FI, indicating that CH₄ fluxes can also contribute to mitigating GHG emissions. However, the current study had some limitations. For instance, we only evaluated direct GHG emissions. Emissions from irrigation power consumption, agricultural machinery fuel oil consumption, and other farm operations are also crucial for the C budget of paddy fields ecosystems [63,64]. All correlative GHG emissions should be considered to evaluate the C budget of paddy field ecosystems after the implementation of water-saving irrigation in future studies. At present, the optimal management pattern we obtained was very consistent with the water-saving and emission reduction policies advocated by the government. Our results also provide data support for the government to determine the appropriate irrigation methods and fertilizer amounts. At the same time, the government should strengthen the promotion to enhance farmers’ awareness of water-saving irrigation. On the one hand, water-saving irrigation reduces farmers’ planting costs; on the other hand, water-saving irrigation reduces the indirect emissions caused by irrigation involved in life cycle assessment (LCA).

5. Conclusions

In Northeast China, we conducted a two-year field experiment to measure Rh, CH₄ fluxes, and NPP and calculated NECB and net GWP. Results showed that single cropping rice paddy fields under each treatment act as net C sinks. The NECB was positive and ranged from 1151 to 2663 kg C ha⁻¹. Compared to FI, single cropping rice paddy fields enhanced C sink intensity after the implementation of II. II increased the NECB through increasing NPP, which exceeded the increased removal of harvest and C mineralized losses. The NECB under II was higher than that of FI and CI by 9–57% and 60–91%, respectively. Compared to FI, the net GWP decreased after the implementation of CI and II. The net GWP under II was also lower than CI when the N fertilization level was 165 kg N ha⁻¹ in 2018 and 2019. These results indicate that II with 165 kg N ha⁻¹ is the optimal management pattern for paddy fields as it can decrease net GWP in Northeast China to combat global climate change. Furthermore, the LCA considering indirect GHG emissions should be applied to evaluate the C of paddy field ecosystems after the implementation of water-saving irrigation in future studies.