The study was conducted from April to August 2024 at Chongqing Modern Agricultural High-tech Park, Jiulongpo District, Chongqing (29°27′N, 106°21′E, elevation 357.7 m). The region’s subtropical monsoon humid climate features abundant water and heat resources, with rains and heat in the same season, an average annual temperature of 16~18°C, the average annual precipitation of 1000~1350 mm (with 70% faling between May and September), and a frost-free period of 340~345 days. During the test period, the average daily atmospheric temperature was 26.8°C, with a high of 41°C and a low of 12°C. The rainfall was 587.4 mm (Figure 1). The typical cropping system involves two to three crops per year. The soil type was purple soil of Shaximiao Formation, characterized by: soil organic matter, 25.2 g·kg^-1; total nitrogen, 3.4 g·kg^-1; ammonium nitrogen, 15.7 mg·kg^-1; nitrate nitrogen, 12.1 mg·kg^-1; pH, 7.1.

This study selected five representative hybrid rice varieties for investigation: ‘Qxiangyou 352’, ‘Shen9you 28’, ‘Shennongyou 228’, ‘Yuxiangyou 8133’, and ‘Shennongyou 446’. Among these, ‘Shen9you 28’, ‘Yuxiangyou 8133’ and’Shennongyou 446’ are medium-early maturing three-line hybrid rice varieties, recommended for cultivation as single-season medium rice in regions below 800 m above sea level in Chongqing. ‘Qxiangyou 352’ is a late-maturing indica-type three-line hybrid rice variety, suitable for planting in late-maturing indica rice zones of Guizhou Province. ‘Shennongyou 228’ is an indica-type three-line hybrid rice variety, adaptable for early or medium rice cultivation in central and northern rice-growing regions of Guangxi, as well as for medium rice production in high-altitude mountainous areas. These varieties were selected due to their widespread promotion and extensive application in rice production systems in Chongqing in recent years.For experiment consisted of 15 plots, each measuring 6 m² (2 m × 3 m), with three replicates per varieties. The planting density was 20 cm × 30 cm, with 2 seeds per hole. The experimental planting mode was for a single-season rice. The field management measures are in line with the local conditions.

The crop management practices included sowing on March 16, 2024, transplanting on April 25, 2024, and uniform harvesting after maturity on August 9, 2024. Fertilizer application consisted of pure nitrogen (150 kg·ha^-1), superphosphate (135 kg·ha^-1 P₂O₅), potassium sulfate (120 kg·ha^-1 K₂O). To ensure maximum fertilizer efficiency, nitrogen fertilizer was applied in three splits in a ratio of 5:2:3. Two days before transplanting, 50% of the total nitrogen amount was applied, plus full doses of P₂O₅ and K₂O. Subsequently, at the tillering stage of rice, 20% of the total nitrogen was applied. Finally, at the heading stage of the rice, the remaining 30% of total nitrogen was applied. Base fertilizer was applied on April 23, tillering on May 17, and ear fertilizer on June 29, 2024.

The water management strategy for the experimental fields adopts a three-stage irrigation plan: flood irrigation is carried out during the seedling stage, and continuous flooding is applied from the tillering stage to the harvest stage, maintaining a water layer of 3~4 cm. One week before harvest, natural drainage is implemented. All experimental fields follow the principle of replenishing water when it is lacking, and the water layer is kept stable at 3~4 cm during irrigation. Before transplanting, the straw of the previous crop is removed and the soil is manually turned over. All concrete-structured experimental fields are covered with waterproof fabric to effectively prevent water and fertilizer leakage, ensuring the reliability of the experimental data.

Greenhouse gas emissions from rice were measured over the entire growth period (106 d) using static box-gas chromatography. Sampling occurred on sunny days (8:30~11:00 am) every 7 days, and postponed on rainy days, until harvest. Gas samples were collected from each plot and the CH₄ and N₂O concentrations were analyzed to calculate emission fluxes.

The static chamber is fabricated from transparent acrylic and is composed of a top box and a bottom base. The top box measures 100 cm in height, and 50 cm in both length and width. It does not contain an internal fan. The bottom base has a height of 15 cm and dimensions of 50 cm in length and width. There is a groove 4 cm deep on the top surface of the bottom base. During the sampling process, a specific quantity of water was maintained in the groove to guarantee a proper seal when the top box is placed on the base. Gas is extracted from the chamber by means of an automatic gas sampling device. Initially, the intake pipe of the automatic gas sampling device is connected to the ventilation hole on one side of the acrylic box. The air within the box is drawn through a rubber tube. The external end of the tube is connected to an aluminum foil gas bag (Delin, Dalian) via a three-way valve to collect the gas samples withdrawn from the box. The automatic gas sampling device is equipped with an internal temperature sensor. The temperature probe is inserted into the opening at the top of the box and then sealed to measure the temperature inside the box. This device is capable of automatically collecting four gas samples. Simultaneously, it measures the temperature during the sampling process. The time interval between each two consecutive samplings is 5 minutes. Sampling commences at 8:30, and the collected gas is stored in aluminum foil gas bags. After each sampling, the top box of the static chamber is removed, leaving only the bottom base in place. This is to ensure the normal growth of rice. Immediately following the gas collection, the aluminum foil gas bags used to gather the experimental gas are sent to the China National Rice Research Institute for concentration detection. The concentrations of CH₄ and N₂O gases are determined using a GC-2010 Plus gas chromatograph (Shimadzu Corporation, Japan).

The calculation formula of greenhouse gas emission flux is as follows (Li et al., 2023) (Equation 1):

Where, F is the emission flux of methane and nitrous oxide, and the unit is mg·(m²·h) ^-1 and ug· (m²·h) ^-1,respectively; ρ is the gas density at standard atmospheric pressure, CH₄ is 0.714 kg·m³, N₂O is 1.25 kg·m³; H is the net height of the sampling box, m; dC/dt is the linear slope of gas concentration change with time during sampling time; T is the average temperature in the box during the sampling process, in °C. The CH₄ and N₂O emission fluxes are represented by three repeated averages.

The formula for calculating CH₄, N₂O emissions in each growth period is as follows (Kim et al., 2019) (Equation 2):

In the formula, R_i represents the mean of the emission of two adjacent measurements, D_i represents the interval of two adjacent measurement dates, d. The total CH₄ emission in the whole growth period was the sum of emissions in each growth period.

The warming potential of CH₄ and N₂O emissions in rice fields (GWP, GWP of CO₂ is 1) was calculated according to the conversion coefficients of 28 and 265 on a 100-year time scale, respectively, and the formula is as follows (Guo et al., 2023) (Equation 3):

Where, GWP represents the warming potential (t CO₂e·ha^-1); CH₄ and N₂O represent the total emission of CH₄ and N₂O gas during rice cultivation (t·ha^-1), respectively.

Greenhouse gas intensity (GHGI) is related to rice yield, and the formula is as follows (Haque et al., 2016) (Equation 4):

Where: GHGI is greenhouse gas emission intensity, t CO₂e· t^-1; Y is the yield of rice, t·ha^-1.

A total of 17 field samples were collected during rice growth period, yieldiing 255 theoretical CH₄ and N₂O emission fluxes. However, not all concentration values resulted in valid emission flux calculations. Following established data processing methods (Lu et al., 2000; Zou et al., 2004), only concentration data with a linear regression≥0.9 were considered valid. After analysis, 231 CH₄ and 225 N₂O emission fluxes were obtained, representing effective rates were 90.6% and 88.2%, respectively.

Plant samples were collected at four critical growth stages of rice: tillering stage, heading stage, full heading stage and filling stage. Three rice plants with uniform growth were selected from each plot. A soil block of 30 cm×30 cm was dug around the root-stem junction of each plant and placed in a sieve. Then, the large soil clumps were washed clean with running water at the edge of the field. Subsequently, the roots of the rice plants were thoroughly rinsed with a high-pressure water gun to ensure that all adhering soil particles were removed as much as possible. After the initial field cleaning, the samples were brought back to the laboratory and further washed with pure water until no sediment remained on the roots. Then, the measurements of plant height, tiller number, leaf area index, aboveground dry biomass, root length, root volume, root porosity, underground dry biomass and root oxidation activity were conducted. The leaf area index (LAI) was determined using the length-width coefficient method (Amanullah and Inamullah, 2016). The longest root length and plant height were measured with a ruler. For measuring the root volume, the water displacement method was adopted. Given the actual size of the rice roots, a graduated cylinder of appropriate capacity was carefully chosen. It was essential that the graduated cylinder could not only fully enclose the rice roots but also provide sufficient space for accurate water-level readings. Initially, an appropriate volume of water was added to the graduated cylinder, and the water-level reading was precisely recorded. The cleaned and surface-dried rice roots were then carefully placed into the graduated cylinder, ensuring that the entire root system was completely submerged in water to avoid floating or entanglement, as these factors could introduce errors in the measurement. A glass rod or other suitable instrument was used to gently stir or press the roots, facilitating their complete submersion to the bottom of the graduated cylinder and expelling any adhering air bubbles. Once the water level in the graduated cylinder had stabilized, the water-level reading was accurately recorded once more. The root volume was then calculated as the difference between the two water-level readings. Regarding the determination of aboveground and underground dry biomass, the samples were first placed in a drying oven at 105°C for 30 min to halt the photosynthetic process, and then dried at 80°C until a constant weight was achieved. Root porosity was measured according to the method of Colmer (2003). The root oxidation activity (ROA) is determined through measuring the oxidation state of α-naphthylamine (α-NA). The detailed methodology can be found in the study conducted by Zhong (2017).

At the maturity stage of rice, three hills of rice were randomly sampled from each plot of the tested rice varieties. After harvest, the rice was manually threshed, and then the plump grains were effectively separated from the empty or shrivelled grains through a flotation method. Subsequently, the two types of grains were dried and counted separately. For the plump grains, the total weight of the entire sample was first measured. Then, a 50 g subsample was randomly selected for a detailed grain count. Regarding the empty or shrivelled grains, their actual quantity was directly counted.

The total number of grains was derived using the total weight of the sample and the calculated thousand-grain weight:

Total number of grains= (Total weight of the sample/Thousand-grain weight of the sample) × 1000.

The data calculation and chart drawing were conducted using WPS Office and Origin 2021 software. Subsequently, a one - way analysis of variance was carried out using IBM SPSS 25.0 software, and multiple comparisons were performed through Duncan’s test (P < 0.05). In this study, the methane and nitrous oxide emissions as well as the rice yield under different treatments were analyzed. Additionally, a principal component analysis (PCA) was implemented on the rice plants and greenhouse gas emissions. Moreover, the Pearson correlation coefficient was also calculated to assess the relationships between the selected factors and methane, nitrous oxide emissions, and rice yield.

Throughout the rice growth cycle, CH₄ emissions displayed a single-peak pattern, with the primary peak occurring during the heading to full-heading stage, followed by a gradual decline. However, the application of tillering fertilizer induced a secondary emission peak during the tillering stage (Figure 2A). At the tillering-stage peak, CH₄ emission fluxes among varieties ranged from 16.4 mg·m^-2·h^-1 to 28.0 mg·m^-2·h^-1, ranked in descending order as: ‘Shen9you 28’>‘Qxiangyou 352’>‘Shennongyou 446’>‘Shennongyou 228’>‘Yuxiangyou 8133’. ‘Shen9you 28’ showed the highest peak (28.0 mg·m^-2·h^-1), exceeding ‘Qxiangyou 352’ (19.3 mg·m^-2·h^-1) by 45.1% and ‘Yuxiangyou 8133’ (16.4 mg·m^-2·h^-1) by 70.7%. The secondary emission peak—occurring later in the growth cycle—ranged from 26.7 mg·m^-2·h^-1 to 48.5 mg·m^-2·h^-1, with varieties ranked as:’Shennongyou 446’>‘Shen9you 28’>‘Shennongyou 228’>‘Yuxiangyou 8133’>‘Qxiangyou 352’. ‘Shennongyou 446’ reached the highest value (48.5 mg·m^-2·h^-1), which was 10.2% higher than Shen9you 28 (44.0 mg·m^-2·h^-1) and 81.7% greater than Qxiangyou 352 (26.7 mg·m^-2·h^-1).

Cumulative CH₄ emissions varied significantly among rice varieties, ranging from 314.6 kg·ha^-1 to 443.4 kg·ha^-1 (Table 1). ‘Qxiangyou 352’ exhibited the lowest emissions (314.6 kg·ha^-1), which were 17.5% to 29.1% lower than all other cultivars. Emissions ranked in ascending order as follows: ‘Qxiangyou 352’<‘Shennongyou 228’<‘Yuxiangyou 8133’<‘Shennongyou 446’<‘Shen9you 28’. No significant differences were detected among ‘Shennongyou 228’, ‘Yuxiangyou 8133’, and ‘Shennongyou 446’.

Total CH₄ emissions also differed across growth stages (Table 1). The heading to full heading stage was the period with the highest CH₄ emissions and clear varietal differences, where ‘Shennongyou 446’ emitted significantly more CH₄ than the other varieties. The filling to maturity stage had the lowest emissions, yet ‘Shen9you 28’ still showed significantly higher fluxes than all other varieties in this period. No significant differences among varieties were found in the other growth stages.

As can be observed from Figure 2B, the N₂O emission patterns of different rice varieties were fundamentally similar. During post-transplant growth, N₂O flux remained low across all varieties, with periods of net uptake observed at various intervals. Emission curves exhibited fluctuating characteristics, with no significant differences between peak values. Toward the end of the growth period, following field drainage and drying, distinct N₂O emission peaks occurred in all varieties. Peak emissions ranged from 28.3 to 50.9 μg·m²·h^-1, ranked in descending order as:’Yuxiangyou 8133’>‘Shen9you 28’>‘Shennongyou 446’>‘Shennongyou 228’>‘Qxiangyou 352’. Among these, ‘Yuxiangyou 8133’ showed the highest peak (50.9 μg·m²·h^-1), exceeding the second-highest emitter, ‘Shen9you 28’ (46.3 μg·m²·h^-1), by approximately 9.9%. Compared to the lowest peak emitter, ‘Qxiangyou 352’ (28.3 μg·m²·h^-1), ‘Yuxiangyou 8133’s emission rate was approximately 80.0% higher.

Cumulative N₂O emissions across rice varieties ranged from -0.049 kg·ha^-1 to 0.165 kg·ha^-1, with no significant differences among varieties (Table 2). Emissions ranked in ascending order as follows:’Yuxiangyou 8133’<‘Shennongyou 446’<‘Qxiangyou 352’<‘Shennongyou 228’<‘Shen9you 28’. ‘Yuxiangyou 8133’ showed the lowest cumulative emission (-0.049 kg·ha^-1), indicating net uptake, while ‘Shen9you 28’ exhibited the highest (0.165 kg·ha^-1).

Significant temporal variations in N₂O fluxes were observed across growth stages (Table 2). From tillering to heading, all varieties consistently consumed N₂O, showing negative emission values. Notably, significant varietal differences emerged during the full heading to filling stage and the filling to maturity stage. In the latter stage, ‘Shen9you 28’ exhibited significantly higher N₂O emissions than all other varieties.

The yields and their constituent factors varied significantly among the rice cultivars tested (Table 3). ‘Yuxiangyou 8133’ produced the highest number of effective panicles per hectare (2.9×10⁶ ha^-1), exceeding the other varieties by 7.4%~11.5%, while ‘Shennongyou 228’ and ‘Shen9you 28’ had the lowest (2.6×10⁶ ha^-1). ‘Yuxiangyou 8133’ also exhibited the highest thousand-grain weight, surpassing other cultivars by 14.5%~24.6%. In contrast, ‘Qxiangyou 352’ had the highest number of filled grains per panicle (200.2 grains), which was 17.0%~60.8% greater than that of the other varieties. ‘Shen9you 28’ achieved the highest seed setting rate (85.4%), whereas ‘Yuxiangyou 8133’ had the lowest (75.4%). No significant difference in seed setting rate was observed between ‘Qxiangyou 352’ and ‘Shennongyou 446’. Grain yield ranged from 7.6 t·ha^-1 to 11.2 t·ha^-1. Owing to the integrated contributions of panicle number, grain weight, filled grains per panicle, and seed setting rate, ‘Qxiangyou 352’ significantly outyielded all other varieties, with a yield advantage of 23.1%~47.4%.

As presented in (Figure 3), the GWP of each rice variety spans from 8.8~12.4 t CO₂e·ha^-1. In ascending order, they are: ‘Qxiangyou 352’<‘Shennongyou 228’<‘Yuxiangyou 8133’<‘Shennongyou 446’<‘Shen9you 28’. Specifically, ‘Qxiangyou 352’ exhibits the lowest GWP, measuring 8.8 t CO₂e·ha^-1, which is markedly lower than that of the other rice varieties. Conversely, there is no statistically significant difference in GWP among the remaining rice varieties. Notably, ‘Shen9you 28’ has the highest GWP, reaching 12.4 t CO₂e·ha^-1, which is approximately 6.9%~40.9% higher compared to the other varieties.

Regarding the GHG, it ranges from 0.8~1.5 t CO₂e·t^-1 for each rice variety. Arranged from low to high, they are: ‘Qxiangyou 352’<‘Shennongyou 228’=‘Shen9you 28’<‘Shennongyou 446’<‘Yuxiangyou 8133’. ‘Qxiangyou 352’ has the lowest GHGI, being significantly lower than that of the other rice varieties by 42.9%~46.7%. Additionally, no significant difference in GHGI is observed among ‘Shennongyou 446’, ‘Shennongyou 228’, ‘Shen9you 28’, and ‘Yuxiangyou 8133’. However, ‘Yuxiangyou 8133’ has the highest GHGI, which is 7.1%~15.4% higher than that of ‘Shennongyou 446’, ‘Shennongyou 228’, and ‘Shen9you 28’.

Significant variations in aboveground physiological traits were observed among the rice varieties (Figure 4). Plant height increased rapidly during the early growth stages and stabilized thereafter. At the booting stage, ‘Qxiangyou 352’ was significantly taller than the other varieties. By the filling stage, ‘Shen9you 28’ exhibited the maximum plant height. Tiller number increased markedly from the tillering to the heading stage before stabilizing. At the tillering stage, ‘Shennongyou 446’ had a significantly higher number of tillers, whereas ‘Shen9you 28’ showed the lowest. No significant differences among varieties were detected in later growth stages. The LAI increased continuously through the tillering, heading, and booting stages, then plateaued. Although no significant differences were observed among varieties at any stage, ‘Shen9you 28’ showed a relatively higher LAI at tillering, while ‘Qxiangyou 352’ and ‘Yuxiangyou 8133’ led at the heading stage. Aboveground dry biomass accumulated consistently across all four growth stages. At the heading stage, ‘Shennongyou 446’ had significantly higher biomass than the other varieties. At the booting and filling stages, ‘Yuxiangyou 8133’ displayed numerically greater biomass, though differences were not statistically significant among varieties including ‘Shennongyou 446’, ‘Shennongyou 228’, and ‘Yuxiangyou 8133’.

As shown in Figure 5, root length increased during early growth stages and stabilized by later stages, with no significant differences observed among varieties across the tillering, heading, full-heading, and filling stages. Root volume increased continuously throughout the growth period. At tillering, ‘Qxiangyou 352’ exhibited significantly greater root volume than the other varieties. At heading, ‘Qxiangyou 352’, ‘Shennongyou 446’, and ‘Shennongyou 228’ all had significantly larger root volume than ‘Shen9you 28’ and ‘Yuxiangyou 8133’. At both full-heading and filling stages, ‘Qxiangyou 352’ again showed significantly greater root volume compared to the other varieties. No significant differences in root porosity were detected among varieties at tillering. From the heading stage onward, no significant differences were observed among ‘Qxiangyou 352’, ‘Shennongyou 446’, ‘Shennongyou 228’, and ‘Yuxiangyou 8133’, although ‘Shennongyou 446’ consistently exhibited relatively higher values. Below-ground dry biomass increased across all growth stages. At tillering, ‘Shennongyou 446’ and ‘Shennongyou 228’ had significantly higher biomass than the other varieties. At heading, both ‘Qxiangyou 352’ and ‘Shennongyou 446’ showed significantly greater biomass than the remaining varieties. By the filling stage, no significant differences among varieties were detected. ROA showed an initial increase followed by a decrease. No significant varietal differences were observed until the filling stage, during which ‘Qxiangyou 352’ and ‘Shen9you 28’ exhibited significantly higher ROA than the other varieties, with ‘Qxiangyou 352’ being the highest.

Correlation analysis (Figure 6) indicated that CH₄ flux rates during both the heading and full-growth stages of the rice varieties were highly significantly positively correlated with key growth indicators measured in these periods. Specifically, CH₄ flux showed significant positive correlations with plant height, LAI, and aboveground dry weight. Concurrently, CH₄ flux was highly significantly negatively correlated with ROA during these growth stages. Moreover, N₂O flux during these stages was also highly significantly positively correlated with plant height, LAI, root volume, and aboveground dry weight, while being significantly negatively correlated with ROA. In terms of cumulative emissions, cumulative CH₄ emissions exhibited a highly significant positive correlation with tiller number and a significant positive correlation with aboveground dry biomass. Cumulative N₂O emissions were positively correlated with LAI, root length, root volume, and root porosity. Furthermore, grain yield was significantly positively correlated with root porosity and belowground dry weight during this growth period.

Principal component analysis (PCA) revealed complex relationships between rice plant traits and greenhouse gas emissions (Figure 7). In the aboveground compartment, PC1 and PC2 accounted for 29.6% and 18.6% of the total variance, respectively. The ordination plot indicated that plant height and LAI were positively associated with cumulative CH₄ emissions, with samples distributed along the positive direction of PC1 suggesting that taller plants may be linked to higher CH₄ release. For the belowground compartment, PC1 and PC2 explained 35.0% and 19.5% of the total variance, respectively. Root length and root porosity were correlated with both CH₄ and N₂O emissions. Patterns in sample distribution revealed intricate associations between root morphological traits and greenhouse gas fluxes, where root volume and belowground biomass exhibited negative correlations with N₂O emissions along the negative PC1 axis.

In this research, the seasonal variations of CH₄ emission fluxes among different rice varieties demonstrated remarkable similarities. All five rice varieties under investigation exhibited a single-peak pattern in CH₄ emissions. The peak occurred during the heading-booting stage, after which the CH₄ emission fluxes showed a steadily declining trend (Figure 2A), which is consistent with the research by Hang et al. (2022). Notably, Jiang et al. (2013) reported two peaks in the rice growth period, one during the tillering stage and the other during the heading stage, which diverged from the findings of this study. The production and oxidation of CH₄ are intricately regulated by methanogenic bacteria and methane-oxidizing bacteria. These microbial populations are highly sensitive to soil temperature, which, in turn, influences soil CH₄ emissions by shaping the microbial community structure and function (Mohanty et al., 2007). In Chongqing, the early rice transplanting time results in relatively low soil temperatures and reduced soil biological activity during the early growth stage of rice. The slow oxygen consumption rate under such conditions is not conducive to the proliferation and activity of methane-producing bacteria. During the heading-booting stage, rice plants generate a substantial amount of root exudates and shed leaves. These provide a rich source of substrates for methanogenic bacteria, leading to an increase in the abundance of methanogenic bacteria and subsequently promoting CH₄ emissions. As the rice reaches the mature stage, the abundance of methanogenic bacteria gradually diminishes, which corresponds to a decline in CH₄ emission level (Fernandez-Baca et al., 2021). Furthermore, during the period from the heading stage to the full heading stage, the root volume, root dry weight, and ROA of ‘Qxiangyou 352’ were all greater than those of other rice varieties. Rice varieties characterized by well-developed root systems and strong oxygen-secreting capabilities can effectively promote CH₄ oxidation and thus reduce CH₄ production (Qi et al., 2024). Consequently, the root system of rice is a crucial factor contributing to the disparities in CH₄ emissions among different rice varieties.

All rice varieties showed a certain peak in N₂O emissions after the paddy fields were drained and sun-dried. This might be attributed to the long-term waterlogged conditions in the experimental fields. As nitrification and denitrification processes are mainly active during the soil’s wet-dry alternation stage, the N₂O emissions from continuously waterlogged paddy fields usually remain at a low level or even negative throughout the growing season (Tokida et al., 2010; Perry et al., 2022). Moreover, under flooded conditions, the diffusion of N₂O is inhibited, facilitating its further reduction to N₂ via denitrification. This also explains the consistently low N₂O emissions observed in the present study (Lin et al., 2017). Notably, significant alterations in soil moisture can expedite the nitrification and denitrification rates (Xiong et al., 2007), this leads to an emission peak after drainage. During the flooding period, due to oxygen deficiency, the N₂O release is minimal. Conversely, the drainage of the paddy field enhances soil aeration, thereby triggering a release peak. Similar findings have been reported by previous scholars (Xiong et al., 2002; Xu et al., 1997).

In this study, the grain yield of the rice varieties under test spanned from 7.6 to 11.2 t·ha^-1. Among them, ‘Qxiangyou 352’ exhibited the highest yield. This could be largely attributed to its remarkable performance in terms of the number of effective panicles, the number of grains per panicle, the thousand-grain weight, and the seed setting rate. Notably, this variety had the highest thousand-grain weight and seed setting rate among all the tested varieties. Furthermore, the yield differences were significantly associated with phenotypic traits including plant height, leaf area index, and root characteristics.

Plant height is a crucial factor influencing rice yield (Badshah et al., 2014). A body of research has demonstrated that a proper increase in plant height can contribute to enhanced yield (Quan et al., 2024; Fernández et al., 2018). Specifically, within a certain range, rice yield shows a positive correlation with the increasing in plant height. The LAI of rice directly impacts the light interception capacity of the biological population, thereby affecting the synthesis of photosynthetic products in rice (Wang et al., 2025). As a result, the LAI of rice is of great significance in the production and accumulation of dry matter as well as the formation of yield. Notably, in this particular study, no significant differences in LAI were observed among various rice varieties. This finding suggests that LAI might not be the primary determinant of rice yield. Furthermore, yield variations among different rice varieties are closely associated with root characteristics. Roots, being the vital organs for rice to absorb water and nutrients (Zhu et al., 2025), their development level and activity directly influence the growth state and ultimate yield of rice. Liu et al. (2020) indicated that the characteristics of high root activity (such as root dry weight and root oxidative activity) are of great significance in promoting the formation of large panicles in super-high-yield hybrid rice. Specifically, the increase in root activity facilitates the absorption of soil nitrogen. This contributes to biomass accumulation and enhances nitrogen utilization efficiency, ultimately boosting rice yield (Zhu et al., 2022). In this study, among all the tested rice varieties, the ‘Qxiangyou 352’ variety exhibited the greatest plant height, along with a higher leaf area index and more favorable root characteristics during each growth stage. These combined advantages firmly established a solid foundation for its highest yield performance. Moreover, Denier van der Gon et al. (2002) showed that a negative correlation exists between the CH₄ emission flux in paddy fields and the rice yield. Jiang et al. (2017) likewise discovered that high-yield rice varieties generally exhibit lower CH₄ emissions. This is consistent with our research findings.

This study demonstrates that the morphological and physiological traits of rice exert a significant influence on CH₄ and N₂O emissions from paddy fields. Specifically, rice varieties characterized by greater plant height, LAI, and higher aboveground/root biomass tend to exhibit elevated CH₄ and N₂O emissions. In contrast, enhanced ROA in rice correlates with a reduction in the emissions of these two greenhouse gases. This observation implies that robust ROA fosters an oxidative microenvironment in the rhizosphere, thereby facilitating CH₄ oxidation and suppressing denitrification processes—ultimately mitigating net GHG emissions (Guan et al., 2024; Qin et al., 2015; Qi et al., 2024). The rice cultivar ‘Qxiangyou 352’ selected for this study possesses well-developed root systems and high aboveground biomass, with ROA significantly superior to other tested varieties. This trait may constitute a key mechanism underlying its ability to maintain high yields while achieving the lowest cumulative CH₄ emissions, comprehensive GWP, and GHGI among the evaluated cultivars. However, the performance of the ROA exhibits variability across different regions. According to the research conducted by Liu et al. (2020) in Hunan province, a significantly positive correlation was identified between ROA and methane emissions. This phenomenon may be attributed to the discrepancies in rice varieties and planting environment. Furthermore, our findings reveal a significant positive correlation between root volume and N₂O emissions, which aligns with the conclusions of (Fu et al., 2011). This relationship can be attributed to the fact that greater root biomass enhances the root system’s capacity for N₂O uptake and translocation, thereby leading to increased N₂O emissions (Fu et al., 2012).