The experiment was conducted in 2022-2023 at the Rice Research Institute of Sichuan Agricultural University in Wenjiang, Chengdu, Sichuan Province, China (30°43′N, 103°51′E), building upon previous experiments (Liu et al., 2019; Yao et al., 2020). The previous crop was wheat, which was naturally air-dried and rolled into small pieces (about 4 cm) at a moisture content of 15% for straw mulching. 4710 kg/hm² of wheat straw (15.0% moisture content) was incorporated into the soil. The soil in the experimental field was sandy loam, with a topsoil depth of 0–20 cm, containing 2.26 g/kg of total N, 20.45 g/kg of organic matter, 115.82 mg/kg of available N, 28.19 mg/kg of available P, and 87.88 mg/kg of available K. The tested rice variety was “Chuankangyou 2115” (provided by Sichuan Agricultural University, a three-line hybrid rice variety with a total growth period of 151.8 d). The microbial agents used were B. subtilis and T. harzianum (both provided by Sichuan Green Microbial Technology Co., Ltd.) at a concentration of 200 × 10⁹ CFU/g.

The application rates of B. subtilis and T. harzianum were based on the method described by (Liu et al., 2018). The experiment used a completely randomized design with five treatments: T1 – no wheat straw return; T2 – return of wheat straw; T3 – return of wheat straw with a combined application of microbial agents (B. subtilis: T. harzianum=50 g: 50 g); T4 – return of wheat straw with a combined application of microbial agents (B. subtilis: T. harzianum=75 g: 25 g); and T5 – return of wheat straw with a combined application of microbial agents (B. subtilis: T. harzianum=25 g: 75 g). Three replicates were conducted in a plot with an area of 4.0 m × 5.0 m = 20.0 m². Wheat straw was fully incorporated into the soil, and one day before transplanting, a microbial agent was applied according to a prescribed ratio, with a total application rate of 50 kg/ha. On April 7, 2022, rice seeds were sown in a seeding field and covered with a film for seedling cultivation. By 15 May, the seedlings were transplanted to the experimental plots and planted at a spacing of 33.3 cm × 16.7 cm between rows and plants, one plant per hole. 12 rows with 29 plants per row were finally planted in each plot. Raised beds (40 cm wide) were constructed between plots and covered with a plastic film. Fertilization and irrigation were consistently managed, and measures were taken to prevent and control diseases, pests, and weeds. The total amount of N applied was 150 kg/ha, with compound fertilizer (Urea- Phosphorus pentoxide - Potassium oxide:15-15-15) applied at a ratio of 5:2:3 for basal, tillering, and panicle fertilization; basal fertilizer was applied 1 d before transplanting, tillering fertilizer was applied 7 d after transplanting, and panicle fertilizer was applied at the booting and heading stages in equal amounts.

Data were processed using Microsoft Excel 2019, and Sigma Plot 14.0 was used for graphical representation. Statistical analysis was performed using SPSS 24.0, one-way analysis of variance and Duncan’s multiple range test were used, and the least significant difference method was employed to determine significant differences between treatment means at P<0.05. Additionally, as the experimental results over two years show a consistent trend, the comprehensive report will primarily focus on the experiments conducted in 2022.

Table 1 shows that all treatments significantly affect rice yield and its components, except for grains per ear and 1000-grain weight in year 2022, and 1000-grain weight in year 2023. The results from the two-year experiment were consistent. Compared with T₁ treatment, the rice yield of T₂, T₃, T₄ and T₅ treatment increased by 3.81%, 17.15%, 10.63% and 10.38% in 2022, and 12.75%, 26.63%, 15.45% and 13.76% in 2023, respectively. T3 treatment, which had the highest yield, was identified as the optimal treatment for combined microbial agent application under wheat straw return conditions. In 2022, from the perspective of yield components, the T₁ treatment had a higher setting rate, which was significantly higher than that of other treatments, but this could not compensate for the lack of effective panicle number and total spikelet number. Compared to the T₁ treatment, the T₂ treatment significantly increased the number of spikelets per panicle by 4.26%; the application of microbial agents in treatments T₃-T₅ could further increase the effective panicle number and total spikelet number, showing a significant increase of 6.97 ~ 11.54% and 9.67 ~ 15.23% compared to T₁, respectively. In 2023, compared to the T₁ treatment, the T₂ treatment significantly increased the effective panicle number by 9.85%; the combined application of microbial agents (T₃-T₅ treatments) significantly increased the effective panicle number by 3.86-33.12% and the total spikelet number by 3.87-23.74%, ensuring an increase in yield. Overall, the two-year data, compared with T₄ and T₅ treatments, T₃ treatment could further increase the number of effective panicle and total spikelets of rice, thus increasing the yield of rice. In this study, T₃ treatment combined with wheat stalk return to the field (Bacillus subtilis: Trichoderma harzia = 50 g: 50 g) was the appropriate treatment.

As shown in Tables 2 and 3 , the root morphology of rice followed a pattern of an initial increase before decreasing during the growth stages, peaking at the heading stage. During the tillering stage, the differences in single-stem root length and root surface area were not significant. However, in other growth stages, significant differences in these two indicators and population root dry weight were observed among the treatments. Compared to the T₁ treatment, the T₂, T₃, T₄, and T₅ treatments increased the single-stem root length (RLS) by 0.88–1.99%, 4.77–10.38%, 5.18–7.77%, and 2.24–8.41%, respectively, from the jointing to the maturing stages. Single-stem root surface area (RAS) increased by 2.12–22.50%, 10.68–23.56%, 3.31–11.06%, and 3.18–15.87% in the T₂, T₃, T₄, and T₅ treatments, respectively, compared with the T₁ treatment; the microbial agent-combined application treatments (T₃–T₅) showed the best performance regarding the RLS and RAS under the T₃ treatment. During the tillering stage, the T₁ treatment had the highest population root dry weight (RWP), significantly higher than the wheat straw return (T₂) and microbial agent-combined application treatments (T₃–T₅). However, from the heading to maturing stages, the microbial agent-combined application treatments (T₃–T₅) were higher than the T₁ treatment, with the T₃ treatment showing the best performance, increasing RWP by 4.80–6.55% compared to the T₁ treatment. In summary, while wheat straw return (T₂) can promote root growth in rice from the jointing stage onwards, the combined application of microbial agents (T₃–T₅) based on straw return demonstrates a more pronounced effect, with the T₃ treatment showing the most significant improvement in root growth.

As shown in Figure 1A , the straw decay rate gradually increased with time, reaching 62.16–74.89% at the rice maturing stage. Compared to the T₂ treatment, the microbial agent-combined application treatments (T₃–T₅) significantly increased the straw decay rate by 12.64–45.97% during all rice growth stages, with the T₅ treatment showing the highest increase in straw decay rate, which was 1.54–10.12% and 5.10–15.75% more than that of the T₃ and T₄ treatments, respectively.

As shown in Figure 1B , soil sucrase activity in all treatments showed a similar trend, reaching a maximum at the tillering stage and then decreasing. Compared to the T₁ treatment, the T₂–T₅ treatments increased soil sucrase activity by 3.33–13.39%, 35.79–64.62%, 21.40–58.21%, and 10.80–47.64%, respectively, with T₃ treatment showing the most significant increase.

As shown in Figure 1C , the soil phosphatase activity in all treatments followed a trend of increasing and then decreasing, reaching a maximum during the jointing stage. Compared to the T₁ treatment, the T₂–T₅ treatments increased soil phosphatase activity by 2.58–8.24%, 19.68–30.89%, 11.67–25.53%, and 1.83–17.33%, respectively, with the T₃ treatment showing the most significant increase, followed by the T₄ treatment; the T₅ treatment showed no significant increase in the soil phosphatase activity.

As shown in Figure 1D , the soil urease activity of all treatments showed a similar trend of increasing and then decreasing, reaching a peak during the jointing stage. Compared to the T₁ treatment, the T₂ treatment significantly increased soil urease activity by 9.80–13.48% from 15 d after transplanting to the tillering stage, while the microbial agent-combined application treatments (T₃–T₅) significantly increased soil urease activity by 19.25–52.21%, 11.76–41.42%, and 8.91–37.25%, respectively, during all rice growth stages. These results indicate that wheat straw return combined with microbial agent application can increase the straw decay rate and soil enzyme activity; the T₅ treatment showed the best effect on straw decay, the T₃ treatment showed the best effect on soil sucrase and phosphatase activities, and the T₃–T₅ treatments showed a significant effect on soil urease activity, with T₃ > T₄ > T₅.

As shown in Figure 2A , except for the heading stage, the total soil N of the T₃ and T₄ treatments was significantly higher than that of the T₁ treatment during all other growth stages, with increases ranging from 8.54% to 16.92% for T₃ and from 7.31% to 12.48% for T₄. There was no significant difference in total soil N between the T₂ and T₅ treatments and the T₁ treatment during all rice growth stages. According to Figure 2B , the T₂ treatment of straw returning was significantly higher than T₁ in TS; the microbial agent-combined application treatments (T₃–T₅) significantly increased soil available nitrogen content compared with T₁ and T₂ at day 15 and HS ( Figure 2B ). T₃ treatment showed the highest content in all periods, which significantly increased by 4.85%-12.46% compared with T₁. These results indicate that combining microbial agents can significantly increase the soil available N content, with T₃ showing the best effect, followed by T₄. The soil available N content showed a decreasing trend during the rice growth stages, and the decrease in the soil available N content of the combined microbial agent application treatments (T₃–T₅) was higher than that of the T₁ treatment, with the T₃ treatment showing the most significant decrease. As shown in Figure 2C , the available phosphorus content in soil treated with microbial agents (T₃-T₅) was significantly higher than that in the T₁ treatment from 15 d after transplanting to TS, with the T₃ treatment showing the greatest increase, which was 6.35 ~ 29.81% higher than the T₁ treatment. Compared to the T₁ treatment without straw return, the soil available phosphorus content in the straw return T₂ treatment significantly increased during the periods 15 days after transplanting, TS, and JS. Within the treatments using microbial agents (T₃-T₅), the T₃ treatment further increased the soil available phosphorus content compared to T₂. The T₅ treatment, which involved the application of microbial agents, showed the lowest soil available phosphorus content during the JS and HS periods. Soil available K ( Figure 2D ) in all treatments showed a decreasing trend during all rice growth stages. Except during the HS period when the available potassium content in the soil of T₁ was higher than that of T₂, the available potassium content in T₂ was higher than that in T₁ during the rest of the periods. However, it is important to note that the differences in soil available potassium content between T₁ and T₂ were not significant in any of the growth stages. The available potassium content in the soil of the microbial agent combined treatment (T₃-T₅) was significantly higher than that in treatments T₁ and T₂ at 15 days after transplanting and during the TS period; the T₃ treatment showed the highest available potassium content in the soil at all growth stages, which was 9.15 ~ 13.14% higher than that in treatments T₁ and T₂.

A total of 3,363,677 valid reads and 93,761 OTUs were obtained by 16S rRNA amplicon sequencing of soil bacteria. The average amplicon length was 420 bp. The number of valid sequences detected in each soil sample exceeded 40,000, and the sparse curve was flat, indicating that the genetic data was sufficient to reasonably estimate bacterial OTUs in the soil samples ( Figure 3 ). The analysis of microbial community diversity index of each treatment found that the Alpha diversity Chao richness index ( Figure 4A ) showed significant differences between T₁ and T₂, T₃, T₄ and T₅ among treatments, while the Shannon diversity index ( Figure 4B ) showed no significant differences among treatments. PCoA was used to analyze the relative contributions of straw returning to field and microorganisms to soil bacterial diversity. From the Figure 4C , it can be observed that there is a certain degree of separation among the soil bacterial communities across different treatments. The community in T₁ (no straw return to the field) is significantly separated from the other treatments along the PCo1 axis, indicating that the soil bacterial community composition without straw return differs greatly from those with straw return. Additionally, T₂ (straw return) compared to treatments with different proportions of microbial inoculants (T₃-T₅) shows that the latter are closer to each other but also display some degree of separation, especially in the T₄ treatment, which was significantly different from T₃ and T₅ in that it was biased more upwards in the PCo2. Thus, this suggests that the application of microbial agents alters the soil bacterial community and that different microbial agent ratios do alter the bacterial community of the soil. The proximity or overlap of T₃, T₄, and T₅, which involve different proportions of microbial inoculants, in the PCoA plot may indicate that although the proportions of the inoculants differ, the trend in their impact on the bacterial community might have certain similarities. Interestingly, the distance between T₃ and T₄ treatments and T₂ (straw return) is closer, whereas T₅ is further from T₂, indicating that T₅ may have a more distinct microbial community change.

Figure 5A shows that the top 10 bacterial phyla include Proteobacteria, Acidobacteriota, Chloroflexi, Actinobacteriota, Gemmatimonadota, Chloroflexi Desulfobacterota, Myxococcota, Nitrospirota, Methylomirabilota, and Firmicutes. In Figure 5A , the average proportion of total soil bacteria among different treatments was 85.56%. Some bacterial groups such as Proteobacteria, Chloroflexi, and Myxococcota maintained stable abundance at the phylum level, but the relative abundance of Actinobacteriota in treatment T₃ was higher than the other treatments, and reached significant levels with treatments T₁, T₂ and T₄.

To further compare the differences in genera composition between samples and illustrate the trends in species abundance distribution, a heatmap analysis was conducted. A clustered heatmap was drawn using the abundance data of the top 20 genera in terms of average abundance ( Figure 5B ). Soil bacteria under different treatments exhibited different abundances at the genus level. Compared to T₁, treatment T₂ had higher abundances of 4-29-1 and Anaeromyxobacter; in treatments T₃-T₅, where microbial agents were applied, the abundances of Defluviicoccus and MBNT15 were relatively higher than in T₁ and T₂; compared to T₁ and T₂, treatment T₃ had a higher abundance of MBNT15, Defluviicoccus and Rokubacteriales.

There was no significant correlation between RLS and RAS and yield components at TS, while there was significant negative correlation between RWP and effective panicle, spikelets number and total yield ( Figure 6A ). However, RLS, RAS and RWP were positively correlated with effective panicle, spikelet number and yield at maturity stages. In addition, RLS was positively correlated with effective panicle number, spikelets number and yield during the whole growth period. In conclusion, root quality and morphological traits significantly affected the increase of effective panicle number and effective spikelets after wheat straw was returned to the field, which ensured the increase of yield.

SDR is significantly positively correlated with effective panicle, spikelets number, and yield during all stages ( Figure 6B ). Figure 6B also shows that during the tillering and jointing stages, SS, SU, and SP are significantly positively correlated with the effective panicles and spikelets number and yield. As shown in Figure 6C , during the jointing and maturity stages, STN is highly significantly positively correlated with the effective panicles and spikelets number; SAN and SAK are significantly positively correlated with the effective panicles, the spikelets number, and yield at all stages; SAP is mainly significantly positively correlated with the effective panicles, spikelets per panicle, spikelets number, and yield during the heading stage. In Figure 6D , during the heading stage, soil microorganisms at the phylum level are mostly positively correlated with yield, with MBNT15, Rokubacteriales, and Latescibacterota genera showing significant positive correlations with the spikelets per panicle, spikelets number, and yield.

In our study, the combined application of microbial agents under straw returning can significantly enhance the root growth, straw decomposition, soil available nutrients from jointing to heading stage of rice, and improve soil microbial community at heading stage of rice, contributing to a significant increase in the number of effective panicles and total spikelets, thereby increasing rice yield.

Crop straw is a valuable and recyclable biological resource, rich in cellulose, lignin, N, P, K, and trace elements (Mishra et al., 2023). The decay of crop straw is a complex biochemical process in which microorganisms play important roles. Extensive studies have been conducted on the microbial decay of straw. Liu et al. (2016) showed that microbial agents accelerated the decay rate of wheat straw in the soil and influenced its final mineralization degree. Mehmood et al. (2020) and Moon (2022) also proved that adjusting the microbial diversity in soil by increasing certain microbial members is an effective method to improve the efficacy of soil-beneficial flora, which can improve the decomposition of substances such as straw, promote the removal of harmful substances and improve soil fertility. These studies indicated that microbial agents can significantly promote the decay of crop straw and nutrient release. This is consistent with the results of this study, where straw return to the field (with microbial inoculant application) treatments T₃-T₅ showed significantly enhanced straw decomposition and nutrient release compared to the straw return treatment without inoculant (T₂). However, it is important to note that different ratios of microbial inoculants had varying effects on straw decomposition (Li et al., 2024). Our study indicates that the decomposition rate of wheat straw was highest under the T₅ treatment, with a microbial agent application ratio of B. subtilis (25 g): T. harzianum (75 g). This suggests that a 1:3 ratio might be more suitable for straw decomposition. This may be due to the fact that T. harzianum secretes cellulase, which plays a more critical role in straw decomposition (Li et al., 2016).

This is consistent with the results of this study, where straw return to the field (with microbial agent application) treatments T₃-T₅ showed significantly enhanced straw decomposition and nutrient release compared to the straw return treatment without microbial agent (T₂). However, it is important to note that different ratios of microbial agent had varying effects on straw decomposition. Our study indicates that the decomposition rate of wheat straw was highest under the T₅ treatment, with a microbial agent application ratio of 25 g:75 g. This suggests that a 25:75 ratio might be more suitable for straw decomposition. This may be due to the fact that a significant amount of T. harzianum secretes more cellulase, playing a more crucial role in straw decomposition (Li et al., 2016).

Soil enzymes act as catalysts for biochemical reactions and are closely related to the physicochemical properties, nutrient status, and biological activity of the soil. These enzymes are essential indicators for evaluating soil fertility and environmental quality (Marx et al., 2001; Chen et al., 2018; Humberto et al., 2020; Hu et al., 2022). Zhang et al. (2023) showed that the application of exogenous microbial agents reshaped the rhizosphere bacterial community structure of bananas and improved the activity of soil enzymes, such as soil catalase and dehydrogenase. Notably, different agents have varying effects.

Ma et al. (2023) applied four kinds of bacteria (XF-5C, Clostridium L13, Bacillus licheniformis PB3, Methylbacterium B0021), the four microbial agents significantly increased the activities of soil urease, soil phosphatase and soil peroxidase to varying degrees, XF-5C and L13 significantly increased soil urease activity, while B0021 significantly increased soil phosphatase activity. Liu A. et al. (2022) demonstrated that after the application of Bacillus, the activities of sucrase, urease, phosphatase, and other enzymes in the soil were significantly improved. Gong et al. (2013) also found that inoculating T. harzianum in rice seedbeds can enhance the activity of soil urease, sucrase, and phosphatase. The results of our study are consistent with previous findings, showing that the soil enzyme activity in the straw return treatment T₂ was relatively higher compared to the no straw return treatment T₁. Moreover, upon the application of different ratios of B. subtilis and T. harzianum on top of straw return, the activities of soil sucrase, phosphatase, and urease in all treatment groups were further increased to varying degrees. However, the degree of improvement varied with different ratios, with the best being a B. subtilis to T. harzianum ratio of 1:1 (T₃). The application of microbial agents can increase the activity of soil enzymes, and the increased activity of soil enzymes can provide more nutrients for the growth of microorganisms, thus promoting the reproduction of microorganisms, which can form a kind of benign cycle for soil ecology (Wu et al., 2023a). Therefore, microbial agents can be used to regulate soil microbial diversity. In this experiment, all three ratios of microbial agents improved soil enzyme activity in the rice field, with a B. subtilis: T. harzianum of 50 g: 50 g being the most effective in enhancing soil enzyme activity. This experiment was conducted in the ecological area of Sichuan, with a wheat straw return rate of 4710 kg/hm² and a water content of 15.0% for agent application. However, the suitability of the optimal microbial agent ratio identified in this study for different ecological areas and soil conditions remains to be further validated.

This study showed that applying microbial agents could accelerate straw decay, increase soil nutrients, and promote the growth of rice root systems, thereby improving rice yields. It is noteworthy that the root phenotypes of the wheat straw-returned treatment (T₂) were mostly weaker than those of the non-returned treatment (T₁) during the early stages of rice growth. The reason for this phenomenon may be due to the incomplete decomposition of wheat straw in the early stage affecting the root growth, while the root phenotypes were lower in the straw-returned treatment with the application of microbial agents (T₃-T₅) compared to the wheat straw-returned treatment (T₂), which may be due to the fact that nutrients in the wheat straw stimulate the activity of soil microorganisms (Guo et al., 2015), increasing the nutrient conversion rate of soil microorganisms, which results in microorganisms competing for nutrients with the rice root system to compete for nutrients, temporarily inhibiting the growth of the rice root system (Singh et al., 2004).

In addition, root length, root surface area and root dry weight were higher in the wheat straw-returned treatment (T₂) than in the control (T₁) during the remaining growth period due to nutrient release from the decomposition of wheat straw; and some of the root phenotypes were further increased in the microbial agents applied on top of the wheat straw-returned treatment (T₃-T₅).This is consistent with the results of Xu et al. (2018), indicating that adding microbial agents could promote plant growth, possibly because the beneficial flora in the microbial inoculants improved the microbial community in the soil (Liu L. et al., 2022). The application of microbial agents increased soil enzyme activity, making nutrients in the soil more easily absorbed by rice and promoting nutrient release from straw decay (Wu et al., 2023b), which is beneficial for the growth of the rice root system. The results of this experiment showed that rice yield, effective panicles, spikelet number, single-stem root length, single-stem root surface area, and root dry weight were significantly correlated. This is consistent with the results of Singh et al. (2004), indicating that better root morphology can contribute to the formation of effective panicles and spikelet numbers, thereby benefiting yield.

This study demonstrated that the application of microbial agents in combination wheat straw return could increase the soil nutrient content, especially before the jointing stage of rice growth. Chen et al. (2021) reported that microbial agents can increase the content of available N and P in soil; Wang et al. (2023) found that microbial agents can not only promote the proliferation of beneficial microorganisms in continuous cropping soil but also change the physical and chemical properties of the soil, such as increasing available phosphorus. Furthermore, Wu and Lin (2003) revealed that microorganisms have a unique impact on nutrient transformation and can release nutrients from insoluble mineral substances, as observed with microbial agents like Bacillus spp. and Pseudomonas spp. with phosphate-solubilizing capabilities. This indicates that microbial agents can convert soil-insoluble nutrients into available nutrients that are easily absorbed by plants, thus promoting plant growth, and this finding is similar to the results of this study. In addition, the soil nutrient results of this experiment showed that wheat straw return could significantly increase the available N and P in the soil, which differs from the results of Liu et al. (2018) and may be attributable to the type of straw and the soil environment. Wang (2006) reported that the combination of microbial residue and organic fertilizer resulted in higher effective panicles and grain-setting rates, showing an increase of 9.30% in effective panicles and 16.66% in grain-setting rate compared to conventional fertilization, indicating that microbial residue combined with organic fertilizer can stabilize tillering and promote panicles; which is consistent with the results of this study, where the application of microbial agents increased the number of effective panicles by 6.97–11.54%. The results of this experiment indicated that the T3 treatment (50g of B. subtilis and T. harzianum respectively) yielded the best results, significantly increasing soil nutrient content, improving root dry weight, optimizing root morphology, promoting effective tillering of rice, and consequently increasing yield.

Rhizosphere soil microbial community is one of the indexes to evaluate soil quality. In general, the richer the soil microbial species and the more active the metabolism, the healthier the soil is and the more suitable it is for plant growth (Xue et al., 2011). The role and application of microorganisms in agricultural production have been increasingly recognized. The booting stage of rice is a critical period for its nutritional growth and coincides with the peak demand for water and nutrients. Analyzing bacterial communities in the soil during this stage is essential for understanding the structure of microbial communities in the rhizosphere of rice (Wei et al., 2010). In this study, there were no significant differences in the Simpson index among different treatments, but the Chao1 index was lowest in the T₁ treatment, indicating a significant difference from other treatments. This suggests that straw return and microbial inoculant application can affect the diversity and richness indices of bacterial communities in the rhizosphere of rice, impacting the microbial community structure. The results of the phylum-level richness index shown that the dominant phyla are Proteobacteria, followed by Acidobacteria, Chloroflexi, Actinobacteria, and Firmicutes. Previous research has indicated that Proteobacteria generally thrive in nutrient-rich environments, suggesting that the application of microbial agents has increased the nutrient content in the soil (Solank et al., 2020). Moreover, Acidobacteria, an important group of bacteria in soil, involved in the degradation of plant residues, iron cycling, one-carbon compound metabolism, and photosynthesis, contributing to material cycling and ecological environment construction (Wang et al., 2016). Research by Li et al. (2020) on the combined application of organic fertilizers also showed that the most abundant phyla under different treatments are Proteobacteria, Chloroflexi, and Actinobacteria. Additionally, variations in geographic location, soil types, and cultivation methods can lead to differences in bacterial communities in rice fields. Among several treatments, T₃ had the highest yield and, compared to other treatments, also had the highest abundance of Actinobacteria, which are widely considered biocontrol agents for plants and effective against rice sheath blight (Wang et al., 2012). Heatmap analysis indicates that fertilization methods can influence the abundance of microbial communities in the rhizosphere soil, with differences observed at the genus level among treatments. The genus MBNT15 was dominant in the T₃ treatment, and its role in enhancing rice production requires further research.