The field experiments were conducted in a greenhouse of Gaoqiao town, Ansai district (36°38’ N, 109°16’ E) from 2023 to 2024. The site is situated at an elevation of 1192 m and characterized by a semi-arid temperate monsoonal climate, featuring a mean annual temperature of 8.8 °C and annual precipitation of 540 mm. Rainfall occurs predominantly between June and September, accompanied by high annual evaporation reaching 1800 mm. Sunshine duration totals 2400 hours annually, accumulated temperatures range from 3500 to 4600 °C, and the frost-free period lasts 160–200 days. The dominant soil type is sandy loam (plow layer properties): pH 7.9, organic matter content of 8.7 g kg^-¹, total nitrogen of 0.56 g kg^-¹, alkaline hydrolyzable nitrogen of 58.76 mg kg^-¹, available phosphorus of 13.98 mg kg^-¹, available potassium of 111.34 mg kg^-¹, and average dry bulk density of 1.2 g cm^-³.

The aboveground plant residues of Astragalus sinicus (Chinese milk vetch) were harvested and used as green manure, which the nutrient composition of 2.13–2.56% N 0.86–1.07% P, 1.53–1.89% K, and 12.1–15.4 C/N ratio. Mushroom spent substrate was collected after removing the plastic film from discarded spent mushroom, which include of 1.82–2.21% N, 0.62–0.94% P, 1.27–1.65% K, 10-13 C/N ratio for mushroom residue. The variety of pepper (Capsicum annuumL.) is Yan’an University pure No. 8.

The experiment was carried out in a controlled greenhouse (plastic tunnel structure, 30 m × 8 m, with temperature control at 20-25 °C ensuring applicability to protected cultivation systems. A randomized block design with five treatments involving green manure and mushroom residue applications for two years (2023-2024): CK (control without amendments), GM (3000 kg/ha milk vetch green manure only), MR (7500 kg ha^-1 mushroom residue only), GMS (3000 kg ha^-1 milk vetch plus 3750 kg ha^-1 mushroom residue), and GMMR (3000 kg ha^-1 milk vetch plus 7500 kg ha^-1 mushroom residue). Each treatment was replicated three times in 30 m² plots separated by 1m isolation zones. All amendments were incorporated into the 0–20 cm soil layer using a rotary tiller for homogeneous mixing. Conventional chemical fertilizers were applied at rates of 180 kg N ha^-¹, 100 kg P₂O₅ ha^-¹, and 120 kg K₂O ha^-¹ for all treatments. Pepper were grown under uniform groundwater irrigation at 50 mm per event, with pest and weed management following local practices.

Soil samples were collected in havesting stage from ten points within each treatment plot, where soil blocks were carefully fractured along natural cracks and mix them together to form one soil sample. After removing roots, gravel, and plant or animal debris while sealed in plastic containers. Upon transportation to the laboratory, three subsamples from the same plot were pooled to form three composite samples per treatment. A portion of each composite sample was stored at 4 °C for analyzing soil carbon and nitrogen fractions, while the remainder was air-dried, sieved, and used for determining basic soil physicochemical properties.

Soil pH was measured using a pH meter with a water-to-soil ratio of 5:1. Nitrate nitrogen and ammonium nitrogen were determined by spectrophotometry after KCl extraction and indophenol blue colorimetry, respectively. The content of total nitrogen (TN) was measured using a Kjeldahl nitrogen apparatus. Total phosphorus (TP) was determined by molybdenum-antimony anti-colorimetry after sodium hydroxide fusion. Total potassium (TK) was measured by flame photometry after sodium hydroxide fusion. Alkali-hydrolyzable nitrogen (AN) was assessed using the alkali diffusion method. Available phosphorus (AP) was measured by the sodium bicarbonate extraction-molybdenum antimony anti-colorimetry method, while available potassium (AK) was determined by the ammonium acetate extraction-flame photometry method. Soluble protein, soluble sugar and vitamin C were measured by coomassie brilliant blue assay, anthrone colorimetry, and 2,6-dichlorophenolindophenol titration methods, respectively. Easily oxidizable organic carbon (EOC) was analyzed via potassium permanganate oxidation and colorimetry. Dissolved organic carbon (DOC) was extracted with distilled water and subsequently measured. Microbial biomass carbon (MBC) was assessed using the chloroform fumigation-extraction technique. The content of SOC was determined by the potassium dichromate oxidation-external heating method.

The calculation method for SOC storage (t ha^-1), change in SOC storage (t ha^-1) and soil carbon sequestration rate (t ha^-1 a) using the following equations:

SOC storage for each soil layer (SOC storage) = SOC × BD × h × 0.1.

Change in SOC storage (t ha^-1) =SOC_st − SOC_sk.

Soil carbon sequestration rate (t ha^-1 a)=Change in SOC storage/t.

Where SOC is the soil organic carbon content (g kg^-¹), BD is the soil bulk density (g cm^-³), h is the thickness of the soil layer (cm), SOC_st (t ha^-1) is the SOC storage under treatment conditions, SOC_sk is the SOC storage under CK, t is the duration of the treatment (years).

Descriptive statistical analysis was conducted using IBM SPSS Statistics 22 (IBM Corp., Armonk, N.Y., USA) to evaluate the mean and standard deviation of each parameter. ANOVA was conducted to assess significant differences among treatments in the same layer (at p < 0.05), followed by the Duncan method for pairwise comparisons. All graphs were drawn using Origin 2025 (Origin Lab Corporation, Northampton, MA, USA). Pearson’s correlation analysis was used to reveal the relationships among the measured parameters. A random forest model was employed to reveal the key soil factors influencing pepper yield, yield, soluble protein, soluble sugar, and vitamin C. These relationships were calculated and visualized using R software, with the ggplot2 package for plotting.

The application of green manure and mushroom residue significantly enhanced soil fertility by increasing the contents of SOC, total and available nitrogen, phosphorus, and potassium. Significant differences in these soil properties among the different treatments across both the 0-10 cm and 10-20 cm soil layers (Figure 1). In the topsoil (0-10 cm), the combined GMMR treatment consistently yielded the highest values for all measured indicators. Specifically, compared to the CK treatment, the GMMR treatment significantly increased the contents of total nitrogen (TN), total phosphorus (TP), total potassium (TK), alkaline hydrolyzable nitrogen (AN), available phosphorus (AP), available potassium (AK), ammonium nitrogen (NH₄⁺-N), nitrate nitrogen (NO₃^--N) and soil organic carbon (SOC) by 17.9%, 56.5%, 27.2%, 88.9%, 34.5%, 46.2%, 53.0%, 71.5%, and 26.4%, respectively (Figure 1). Consequently, the ranking for all soil nutrient levels in the 0-10 cm layer was consistently observed as GMMR > GMS > MR > GM > CK. This trend persisted in the subsurface soil (10-20 cm), where the GMMR treatment also demonstrated superior performance, increasing the content of TN, TP, TK, AN, AP, AK, NH₄⁺-N, NO₃^–N, and SOC by 25.4%, 61.5%, 45.4%, 48.7%, 45.1%, 65.3%, 87.7%, 53.9%, and 54.2% compared to CK (Figure 1). Furthermore, as expected, the contents of all soil nutrients exhibited a decreasing trend with increasing soil depth.

The content of soil dissolved organic carbon (DOC), dissolved organic nitrogen (DON) and soil inorganic nitrogen (SIN) varied significantly among treatments in both soil layers (Figures 2a–c). The content of DOC, DON and SIN was consistently higher at the soil depth of 0-10 cm than 10-20 cm, with the order GMMR > GMS > MR > GM > CK in both soil layers (Figures 2a–c). In the 0-10cm layer, the GM, MR, GMS, and GMMR treatments increased DOC by 7.8%, 12.6%, 29.0%, and 38.8%, with increased DON by 3.0%, 12.3%, 21.5%, and 41.3%, which increased SIN by 13.6%, 29.6%, 34.6%, and 63.1% compared to CK. In the 10-20 cm layer, all treatments also significantly enhanced DOC, DON and SIN, which improvements ranging from 9.2%-54.2%, 7.2%-40.4%, and 8.6%-57%.

The effects of green manure and mushroom residue on soil microbial biomass carbon (MBC) and easily oxidizable organic carbon (EOC) are shown in Figures 2d, e. Significant differences in MBC content were observed among treatments, with the overall order were GMMR was highest. The content of MBC in GMS treatments increased by 8.2%, 18.8%, 25.7% than GM, MR, CK treatment in the 0-10 cm, which 4.5%, 32.9%, 31% higher than GM, MR, CK treatment in the 10-20 cm soil layer (Figure 2d). The content of EOC in GM, MR, GMS, and GMMR treatments increased by 5.5%, 17.4%, 19.4%, and 27.4% compared to CK, respectively (Figure 2e). In the 10-20 cm layer, the content of CK were lower by 39.2%, 48.8%, 48.5%, and 50.6% than GM, MR, GMS, and GMMR treatments, respectively (Figure 2e). These results demonstrate that the synergistic application of green manure and mushroom residue (particularly the GMMR treatment) can effectively enhance the relatively active easily oxidizable organic carbon fraction in the soil, contributing positively to soil fertility and carbon pool quality.

The effects of green manure and mushroom residue applications on SOC storage (Figures 3a, b) and changes in SOC storage (Figures 3c, d) in the 0-10 cm and 10-20 cm soil layers are illustrated in Figures 3a–d. SOC storage showed a decreasing trend with the increase in soil depth (Figures 3a–d). Green manure and mushroom residue significantly affected SOC storage in both soil layers. SOC storage in GMMR, GMS, MR, and GM treatments was significantly higher than that under the CK treatment (p < 0.05), while no significant difference was observed between the MR and GM treatments (p>0.05) (Figures 3a, b). The SOC storage in the GMMR, GMS treatment in the 0-10 cm increased by 24.25%, 24.49%, 17.56% and 4.39% compared with the CK treatment, which increased by 54.28%, 31.55%,25.31% and 9.46% compared with the CK treatment in 10-20 cm layers, respectively. Overall, the SOC storage under different treatments followed the order: GMMR > GMS > MR > GM.

Soil carbon sequestration efficiency reflects the conversion degree of unit exogenous organic carbon in soil and is used to characterize the evolution characteristics of SOC. Green manure and mushroom residue applications had significant effects on the SOC sequestration rate in the 0-10 cm and 10-20 cm soil layers (Figures 3e, f). In both soil layers, the carbon sequestration efficiency gradually increased with the increase in the application rate of green manure and mushroom residue. The soil carbon sequestration efficiency of the GMMR, GMS, MR and GM treatments was 1.07 t ha^-1 a, 0.98 t ha^-1 a, 0.703 t ha^-1 a, and 0.17 t ha^-1 a in the 0-10 cm soil layer, respectively, which were 1.7 t ha^-1 a, 0.98 t ha^-1 a, 0.79 t ha^-1 a, and 0.29 t ha^-1 a in the 10-20 cm soil layer.

The application of green manure and mushroom residues significantly influenced crop yield and quality. All treatments exhibited varying degrees of enhancement Compared to CK, with the combined application of green manure and mushroom residues (GMS and GMMR treatment) showed the best performance. Compared to CK, the GMMR and GMS treatment achieved the highest pepper yield, representing a significant increase of 17.55% and 12.52%.The MR and GM treatment increased yield by 8.91% and 3.39% compared to CK, which no significant difference was observed between these two treatments (Table 1).

Regarding quality-related physiological indicators, the highest soluble protein content in the GMS treatment, which represented a 10% increase over CK, and also showed the greatest content of vitamin C, with an increase of 25.84% (Table 1). The GMMR treatment exhibited a content of soluble sugar similar to that of the MR treatment, which was 4.87% higher than CK. Meanwhile, the content of vitamin C under the GMS treatment was notably higher than that in CK, with an increase of 25.13% (Table 1). Overall, the effects of different treatments on the measured indicators followed a consistent trend was GMMR ≥ GMS > MR > GM > CK. This indicates that the combined application of green manure and mushroom residues not only significantly improves pepper yield and quality but also synergistically enhances plant physiological metabolism and quality formation. Among the treatments, GMMR showed greater advantages in yield enhancement, while GMS performed better in certain quality-related indicators.

According to the spearman correlation analysis (Figure 4a), there are varying degrees of significant correlations between soil nutrients, carbon-nitrogen components, and organic carbon storage under treatments with fungal residue and green manure and the soluble protein, soluble sugar, VC, and yield. Soil nutrient indicators (AP, AK, NH₄⁺-N) also exhibit highly significant strong positive correlations with yield, soluble protein, soluble sugar, and VC. This indicated that the supply of available soil nutrients is fundamental to ensuring both high quality and high pepper yield. The content of SOC, DOC, MBC and soluble protein, soluble sugar, VC, and yield all show extremely significant positive correlations (p ≤ 0.001, Figure 4a), with mostly moderate to strong correlation strengths (Mantel’s r≥0.25). This demonstrated that the richness and activity of the soil organic carbon storage directly promote the formation of pepper quality traits and the increase in yield (Figure 4a).

The random forest model showed the soil factors explained 97.4%, 88.9%, 93.8%, and 96.9% and of the variability in the yield, soluble protein, soluble sugar, and VC, respectively (Figure 4b), whereas MBC is key soil factors affecting yield and soluble sugar, and NO₃⁺-N is key soil factors affecting soluble protein. Addition, AP is key soil factors affecting VC (p < 0.05, Figure 4b).

In this study, GMMR treatment significantly increased TN, TP, TK, AN, AP, AK, NH₄⁺-N, NO₃^–N, and SOC in the 10-20 cm soil layer, with the same ranking as in the 0-10 cm layer (GMMR > GMS > MR > GM > CK). This aligns with Wang et al. (2021) in their study of Rational utilization of Chinese milk vetch improves soil fertility and production in cropping systems, which concluded that leguminous green manure can simultaneously provide active carbon sources and nitrogen fixation while systemically enhancing soil fertility and crop productivity. Additionally, mushroom residue as an organic amendment rich in organic matter and mineral elements, which can establish a more stable nutrient pool and slow-release effect, thereby amplifying the synergistic benefits of combined application (Martín et al., 2023). This study demonstrated that all treatments significantly increased the contents of AP and AK besides for the CK treatment. The content of TN, TP, TK, AN, AP, AK, NH₄+-N, NO₃^–N, and SOC exhibit a gradual decrease with increasing soil depth. This is because the topsoil layer serves as the primary site for organic matter input and microbial activity, while downward migration is restricted (Jobbágy and Jackson, 2000). However, mechanical incorporation methods can still maintain significant soil improvement effects in the 10-20 cm layer. Furthermore, adequate organic inputs enhance aggregate stability and cation exchange capacity, thereby improving nutrient retention and buffering capacity for elements such as phosphorus and potassium (Annabi et al., 2011; Abiven et al., 2009). Consequently, this provides a structural explanation for the sustained increases in the content of TN, TP, TK, AN, AP, AK, NH₄⁺-N, NO₃^–N, and SOC in both soil layers.

Significant differences in the treatments of green manure and mushroom residue application on thecontent of DOC, DON, SIN, MBC, and EOC showed the potential of combined applications of green manureand mushroom residue to improve soil health and fertility. The content of DOC in the soil variedsignificantly among different application treatments. Notably, in the 0-10 cm layer, the GMMR treatment demonstrated the most significant increase (47.69%), while in the 10-20 cm layer, it achieved a 42.11% improvement. This finding supports the research by Ren et al. (2024), which pointed out that the application of organic materials can significantly enhance the content of soluble organic carbon in the soil, thereby benefiting the functionality of soil ecosystems. The content of DON also reached its maximum value under the GMMR treatment, showing increases of 68.42% and 65.00% compared to CK, respectively. This phenomenon aligns with the findings of Yao et al. (2021), which showed that the introduction of organic nitrogen sources, particularly through green manure and mushroom residue, effectively increases the available nitrogen in the soil for plant growth.

The GMMR treatment exhibited the highest content of SIN, with increases of 35.29% and 41.18% in the 0-10 cm and 10-20 cm layers, respectively. This indicates that the combined application significantly enhances the content of SIN readily available for crop uptake, consistent with the findings of Liu et al. (2020), which demonstrated that the combined use of various organic fertilizers can promote the accumulation of inorganic nitrogen in the soil. Similarly, the GMMR treatment showed the best performance in the content of MBC, with an increase of 48.21%in the topsoil layer. This suggests that the application of green manure and mushroom residue can provide a rich source of available carbon for soil microorganisms, enhancing their activity and biomass. According to Feng et al. (2024), microbial activity is closely related to the addition of organic carbon, which further supporting our findings. In the 0-10 cm layer, EOC increased by 46.67% under this treatment. This is particularly important for improving soil fertility and enhancing the quality of the carbon pool, in line with the conclusions of Liptzin et al. (2022), which indicated that easily oxidizable organic carbon components are crucial indicators for assessing soil health.

Integrated application of green manure and mushroom residue, with the GMMR treatment showing the most pronounced effects, systematically enhances soil physicochemical properties and biological activity. These improvements collectively establish a synergistic physical chemical biological basis for increasing both pepper yield and quality. In this study, random forest model and correlation analyses jointly indicated that the AP exhibits the strongest driving relationship with vitamin C, whereas NO₃^–N showed the closest association with soluble protein content (Figure 4). This confirms that the form and intensity of nutrient supply directly regulate secondarymetabolic pathways in plants. Green manure contributes biological nitrogen fixation and readily mineralizable nitrogen sources, while mushroom residue provides slow-release phosphorus, potassium, and other trace elements, collectively established a nutrient supply regime that integrates immediate availability with sustained release (Han et al., 2026; Liang et al., 2024; Li et al., 2020). This integrated nutrient supply regime not only satisfies the substantial nitrogen, phosphorus, and potassium requirements of peppers to support high yield, but also offers targeted nutritional support for the synthesis of quality-related constituents (Lenka et al., 2022). These include vitamin C, which relies on phosphorus-mediated photophosphorylation and reducing equivalents, as well as soluble proteins.

Furthermore, our research found that the significant positive correlations observed among SOC, MBC, DOC, and all yield and quality indicators reveal a deeper regulatory mechanism (Figure 4). The addition of organic materials not only elevates the total carbon pool but also specifically promotes the formation of active carbon fractions (Mustafa et al., 2020; Zhang et al., 2023). Consequently, the highest carbon sequestration efficiency and active carbon content observed under the GMMR treatment not only represent superior carbon storage potential but also signify a more functional and plant-growth-promoting rhizosphere microenvironment.