The experiment was conducted during the year 2020 to 2022 at the International Rice Research Institute South Asia Regional Centre (ISARC), Varanasi (25.30° N – 82.95° E, 83 m above mean sea level), Uttar Pradesh, India ( Figure 1 ) with rice-based cropping systems under different farming scenarios. The mean minimum and maximum temperatures at the experimental site during experimental period was 15.5°C and 36°C, respectively. Annual rainfall recorded ranges between 900 and 1113 mm and more than 70% rainfall occurred during the monsoon season from July to September ( Figure 2 ). The soil at the experimental site is sandy clay loam and it was classified as Typic Natrustalf (US Taxonomy).

The ISARC, Varanasi was selected as it is a site representative of rice-based cropping systems in similar agroecological zones. While the findings provide valuable insights, the results may be limited in their generalizability to other regions with different environmental and soil conditions. We advocate that future research should be conducted across diverse locations to validate these findings and extend their applicability to broader contexts.

The experiment was established in a randomized complete block design in which four farming scenarios constituted the treatments as follows: Conventional farming (Scenario 1) ploughing and use of chemical-based fertilizers and pesticides, Organic farming (using bio-pesticide, farm yard manure and compost) with conventional tillage (CT) for rice and wheat (Scenario 2), Integrated nutrient management (Organic and chemical based combined input, Scenario 3), and Conservation agriculture with zero tillage (ZT) for rice and wheat (Scenario 4). There were three replications of each treatment. While rice and wheat were grown under all the scenarios, in Scenarios 2 and 4, ZT mungbean was grown as a Zaid crop between wheat and rice.

The management practices, viz. tillage, sowing method, fertilizer application, water management, crop geometry and spacing (between the rice, wheat and mungbean system, as provided in Table 1 . It is also same in all the scenarios, 22.5– 22.5– 45 cm), and crop varieties used for growing rice, wheat, and maize under different farming scenarios are described in Tables 1 , 2 . Crop geometry in agronomy refers to the arrangement of the crop plants in each plots, including the shape of the space available for each plant, and the spacing between rows and plants. While in Scenario 1 wheat was sown by broadcasting the seeds and rice was manually transplanted, in Scenario 3 Happy Seeder machine was used for seeding both rice and wheat.

Fertilizer management practices for different crops under the four scenarios were: Scenario 1 – 120 kg Nitrogen (N), 60 kg Phosphorus (P) and 40 kg Pottassium (K) per ha for rice and wheat, Scenario 2 - bio-inputs (BGA, Azolla, Vermivash, Vermicompost, FYM) for both rice (BGA and Azolla only applicable for paddy) and wheat, and Scenario 3 using 75% of RDF along with 25% bio-inputs (bio-liquid formulations: details mentioned in Table 1 ) and Scenario 4 - Site-Specific Nutrient Management (SSNM) and 25% bio-inputs (bio-liquids formulations). The specific quantities for SSNM are detailed in Tables 1 , 2 . For SSNM, 75% RDF (90:45:30 NPK) with 25% from bio-inputs was used. RDF in Scenario 1 is the same as specified for other scenarios, except when combined with bio-inputs.

Water management practices also varied, with Scenarios 1 and 2 using frequent irrigations for rice and wheat and Scenario 3 and 4 using alternate wet and dry (AWD) methods. The exact number of irrigations for each scenario is as follows: Scenario 1: 30 irrigations for rice, 3 for wheat; Scenario 2: 26 irrigations for rice, 3 for wheat, 1 for mungbean; Scenarios 3 and 4: 22 irrigations for rice, 3 for wheat, 1 for mungbean.

The crop varieties used in all scenarios were Arize 6444 Gold for rice, PBW 187 for wheat, and Virat for mungbean. These varieties were selected based on their suitability and high yield potential in Eastern Uttar Pradesh, India.

After two cropping system cycles (2020-2022), soil samples were collected from the surface layer (0-15 cm) randomly within each plot by using an auger. Three samples within a plot were thoroughly mixed to make composite samples. Composite samples were air-dried under shade. One portion of the sample was ground and the whole amount was filtered using 2.0 mm sieve as per the requirement of different parameter analysis.

The initial soil properties (pH, EC, organic carbon, and available N, P, K) were estimated in air-dried samples using standard protocols ( Table 3 ).

At crop maturity, rice, wheat and mungbean crops were harvested manually from 2×2 m² randomly selected three places from each plot for recoding the grain yield. System productivity was calculated on rice equivalent yield (RQY) basis for wheat and mungbean to express the overall impact of treatments. Grain yield was reported at 15% moisture. System productivity (t ha^-1) was computed using the following Equation.

Where, MSP- minimum support price and INR-Indian Rupees.

The Windows-based XLSTAT software program was used to calculate the mean of the data that was analyzed for each parameter and then processed for analysis of variance (ANOVA) to establish the statistical significance of the consequences of various scenarios. The data were subjected to variance analysis to identify relevant effects, and the means were compared using Tukey’s HSD test as a post hoc mean separation test (P 0.05). Bivariate Pearson’s correlation coefficients, regression equations, and PCA were computed for all soil variables, namely pH, EC, SOC, BD, DHA, APA, MBC, fungus, bacteria, actinomycetes, and soil respiration under different farming scenarios.

No human based study was performed in this work. Experiments were as per the IRRI mandate.

The present study investigated the effect of different scenarios on soil chemical properties. The results showed no significant differences in bulk density (BD), pH, EC, and water holding capacity (WHC) among the four farming scenarios. Growing rice and wheat under different farming scenarios exerted a significant impact on SOC, available N, available P and available K ( Table 4 ). The SOC content ranged from 0.49% to 0.59% among the farming scenarios; the highest value was observed in Scenario 2. The SOC stock was the highest in Scenario 4 (13.75 Mg C ha^-1) and the lowest in Scenario 1 (12.65 Mg C ha^-1). The variations in SOC content and stock can be attributed to the management practices in each scenario, such as the type and amount of organic inputs, tillage practices, and cropping systems. The available N was the highest in Scenario 1 (144 kg ha^-1) and the lowest in Scenario 2 (140 kg ha^-1), while available K was highest in Scenario 2 (56 kg ha^-1) and lowest in Scenario 3 (38 kg ha^-1). The available P was highest in Scenario 1 (38.5 kg ha^-1) and lowest in Scenario 3 (12 kg ha^-1).

The differences in available nutrient content among the scenarios can be due to the differences in the soil organic matter content and the nutrient management practices. The findings of this study are consistent with previous studies (Lal, 2004). Magnitude of SOC stocks at any place is decided by various governing parameters, such as, land use pattern, extent of land degradation, soil type, depth profile and total input of biomass C annually. Therefore, practices that increase SOC content, such as conservation tillage, cover cropping, and organic amendments can enhance soil quality and crop productivity. SOC is one of the most important sustainability parameter of soil health and agricultural sustainability that work at the ecosystem interface (Shelef et al., 2018; Mishra et al., 2022b). It also helps quantitatively in mitigating the net rate of greenhouse gas emissions.

The results suggest that management practices that enhance SOC content could improve soil fertility and nutrient availability. Further studies are needed to investigate the long-term effects of different management practices on soil properties and crop productivity.

The measured soil penetration resistance values at different soil depths and scenarios are presented in Figure 3 . The results indicate variations in soil compaction across the different scenarios and depths. At a soil depth of 5 cm, all scenarios exhibited relatively low penetration resistance values, ranging from 0.07 MPa to 0.14 MPa. This suggests minimal compaction at this shallow depth, which could be attributed to reduced external forces or a less compacted topsoil layer. As the depth increased to 10 cm, the soil penetration resistance values generally doubled across all scenarios. This suggests a gradual increase in soil compaction, potentially due to increased weight-bearing and compaction forces exerted on the underlying soil layers. At a depth of 15 cm, significantly higher soil penetration resistance values were observed compared to the shallower depths, ranging from 0.47 MPa to 1.45 MPa.

The data as plotted in Figure 3 suggest the presence of compacted soil layers, which can hinder plants’ root growth and nutrient uptake. The soil penetration resistance values at depths of 20 cm, 25 cm, and 30 cm remained relatively constant across all scenarios, with values ranging from 1.80 MPa to 2.93 MPa. These consistently high resistance values indicate considerable soil compaction at these depths, which may limit root penetration and water movement through the soil profile. The variations in soil penetration resistance across the farming scenarios highlight the impact of different agricultural practices or land use patterns on soil compaction. Further investigations are necessary to identify the specific factors contributing to these differences.

Organic farming systems had a (46%) higher pool of very labile C (CVL) compared to conventional (p=0.05). Soils under rice–wheat-mungbean cropping systems (Scenario 2 & Scenario 4) exhibited larger very labile C pool than in soils under Scenario 1 & Scenario 3. The results showed that the distribution of SOC fractions varied significantly among the four scenarios. In Scenario 1, the easily oxidizable fraction (CVL) had the highest proportion (0.30), followed by moderately oxidizable (CL, 0.08), slowly oxidizable (CLL, 0.04), and non-oxidizable (CNL, 0.07) fractions. In Scenario 2, the proportion of CVL was the highest (0.44), followed by CNL (0.07), CL (0.05), and CLL (0.03). In Scenario 3, the CVL fraction had the highest proportion (0.26), followed by CNL (0.09), CL (0.10), and CLL (0.06). In Scenario 4, the proportion of CVL was highest (0.38), followed by CNL (0.10), CLL (0.03), and CL (0.03) ( Figure 4 ).

This study showed that different management scenarios significantly affected the distribution of SOC fractions in the soil. The organic farming (Scenario 2) had the highest proportion of easily oxidizable SOC, consistent with previous studies showing that organic management practices can increase the amount of labile C in the soil (Chen et al., 2019). This suggests organic management practices can promote SOC turnover and improve soil fertility. The conservation agriculture scenario (Scenario 4) also had a high proportion of easily oxidizable SOC, possibly due to cover crops and reduced tillage, which can enhance SOM (García-Orenes et al., 2013; Mishra et al., 2022b). Although it is generally acknowledged that soil organic matter is primarily concentrated in the top 30 cm of the soil, there is mounting evidence that, despite the concentrations in the subsoil, deeper soil horizons have the ability to sequester significant amounts of SOC. The standard fixed depth of approximately 1 m is typically regarded appropriate in routine soil surveys, which is one factor contributing to the uncertainty. In light of this, it is still unsure how much the worldwide SOC budget has been underestimated.

Further, this research is of relatively short duration (2020-2022). While our research provides important insights into the immediate and short-term effects of various farming practices on soil health and carbon dynamics, these findings may not fully capture the long-term implications. Soil processes, especially those related to carbon sequestration and microbial community shifts, often require extended periods to manifest significant and sustained changes. Future research focused on long-term studies to better understand the enduring effects of these agricultural practices on soil health, carbon storage, and overall ecosystem resilience is needed. This will help to validate the findings presented here, providing a more comprehensive understanding and benefits of sustainable farming practices over time.

Significant coefficients of correlations (p ≤ 0.01 and p ≤ 0.05) were observed among most soil physicochemical and biological properties, irrespective of different farming scenarios. Based on the correlation coefficients listed in Figure 5 , it can be inferred that there are strong correlations between some of the variables.

There is a strong negative correlation (- 0.577) between BD and SOC (%). This suggests that as BD increases, SOC decreases, and vice versa. Similarly, a strong negative correlation exists between SOC and soil pH (- 0.485) indicating that as pH decreases, SOC tends to increase. There is also a strong positive correlation (0.889) between available P (Av. P) and available K (Av. K). This suggests that as the concentration of one nutrient increases, the concentration of the other also tends to increase.

There are strong positive correlations between some soil biological properties, such as bacteria and actinomycetes (r = 0.660), and between bacteria and fungi counts (r = 0.763). These correlations suggest that these microorganisms tend to coexist and influence each other’s populations in the soil. Additionally, there is a strong positive correlation (0.662) between total glomalin-related soil protein (TGRSP) and soil respiration (SR), which suggests that TGRSP may be contributing to soil respiration the ecosystem ( Figure 5 ).

Table 7 presents the crop yield data from different agricultural practices during the 2020-21 and 2021-22 periods. In the Kharif season of 2020-21, the highest rice yields were obtained from Scenario 3 (4.55 ± 0.20 t ha^-1), closely followed by Scenario 1 (4.35 ± 0.26 t ha^-1), while Scenario 2 (2.88 ± 0.21 t ha^-1) exhibited the lowest yield. Similarly, during the Rabi season, Scenario 3 (3.66 ± 0.19 t ha^-1) and Scenario 1 (4.03 ± 0.17 t ha^-1) showed comparable yields, while Scenario 2 (3.00 ± 0.17 t ha^-1) lagged. When considering the system yield (system productivity), which represents the overall productivity across all seasons, Scenario 4 (10.95 ± 0.49 t ha^-1) emerged as the most productive approach, significantly outperforming the other scenarios. It was followed by Scenario 1 (8.54 ± 0.44 t ha^-1) and Scenario 3 (8.35 ± 0.40 t ha^-1), while Scenario 2 (7.08 ± 0.42 t ha^-1) exhibited the lowest overall yield.

These results suggest that conservation agriculture, characterized by practices such as reduced tillage and soil cover, can significantly enhance crop productivity (Kumar et al., 2021). The observed variations in crop yields can be attributed to multiple factors, including differences in soil health, nutrient availability, pest management (Mishra et al., 2015; Kumar et al., 2022), and overall management practices associated with each farming approach.

Use of organic fertilizers as opposed to chemical fertilizers can effectively mitigate soil erosion and minimize the potential for aquatic ecotoxicity. Moreover, the implementation of organic farming practices has been found to have a positive impact on soil biodiversity and the presence of both macro and microorganisms. This, in turn, leads to increased income per hectare. However, it is important to note that the yield of crops may be lower when employing OF practices, with a reduction of up to 22%. As per the existing literature, empirical evidence suggests that application of organic fertilizers may result in a decrease in crop yields per hectare. Notably, studies have consistently demonstrated reductions of approximately 20-25% and 50% in yields when comparing OF to conventional fertilizers showed the potential benefits associated with the integration of innovative plant breeding technologies, such as CRISPR/Cas9, within the framework of organic farming practices. Whereas, application of chemical fertilizers has been widely recognized for its positive impact on crop production. However, it is important to acknowledge that the use of chemical fertilizers can also have detrimental effects on soil physicochemical properties and subsequently influence the composition and functioning of soil microbiomes.

In the PCA of 21 variables, three principal components were extracted with eigenvalues> 0.9 and these explained 91.18% of the variance. A correlation study (Pearson’s correlation) was performed for all the variables ( Figure 6 ). One of the most significant indicators of changes in soil quality was found to be the biomass of soil microbes (Stenberg, 1999). Correlations exist between microbial activity in the soil, microbial biomass, enzyme activity and SOM contents (Chaer et al., 2009). All soil microbial characteristics significantly (p>0.05) correlated with MBC but not with SOM. Soil MBC was found to be a reliable indicator for determining soil quality. DHA was removed from the MDS to reduce redundancy because it is closely connected to MBC. As a result, the MDS continued to use APA.

Based on PCA, most of the microbiome parameters in organic farming were selected to predict the optimum soil quality by different agricultural practices and qualified as key indicators of soil quality. In the present study, all the indicators except BD, available N, pH, (high and low) labile pool of carbon that were retained in the minimum data set were considered good and when in increasing order, these were scored as “more is better”. In contrast, chemical parameters were scored as “less is better”.