The experimental materials selected for this study include Poa pratensis var. anceps Gaund. cv. ‘Qinghai’ and Puccinellia tenuiflora cv. ‘Tongde.’ The seeds were provided by the Qinghai Academy of Animal Husbandry and Veterinary Sciences.

The experiment consists of three treatments:

The experiment was sown on July 15, 2018, with each plot covering an area of 50 m² and having three replicates. The total sowing rate for each treatment was 22.5 kg·hm^-2 with a row spacing of 30 centimeters. In the mixed sowing treatment, seeds were sown in a 1:1 ratio by seed count. Before sowing, urea was applied at a rate of 75 kg·hm^-2, and diammonium phosphate at a rate of 150 kg·hm^-2 was applied in the same year and the following year after the grasses regrew. Weeds were manually removed once in both the sowing year and the year after the grasses regrew. The experiment was rainfed.

The experimental site is located in the Perennial Forage Germplasm Resource Nursery in Xihai Town, Haibei Prefecture, Qinghai Province, China (36°59′36″N, 100°52′84″E, elevation 3,156 meters above sea level). It falls under a high-altitude continental climate characterized by long cold periods. The annual average temperature is 0.9 °C, with the highest temperature recorded at 30.5 °C and the lowest at -33.8 °C. The cumulative temperature ≥10 °C is 634.5 °C annually. The area receives abundant sunlight with strong solar radiation. The climate exhibits distinct wet and dry seasons, with rainfall and high temperatures occurring simultaneously. The annual average precipitation is 369.1 mm, and there is no absolute frost-free period. The soil type at the site is classified as black calcareous soil. Soil nutrient content is as follows: organic matter 38.35 g·kg^-1, available nitrogen 2.58 mg·kg^-1, available phosphorus 1.36 mg·kg^-1, available potassium 21.69 mg·kg^-1, and soil pH of 8.43.

The yield of P. pratensis and P. tenuiflora was determined during the milk ripening stage in the years 2019 (2nd year), 2020 (3rd year), 2021 (4th year), 2022 (5th year), and 2023 (6th year). The experiment had 3 replicates, with three uniform 1m × 1m = 1m² quadrats selected in each plot. Each treatment was sampled 9 times. The grass in each quadrat was cut at ground level to measure the fresh weight of the pasture. In the mixed sowing treatment, the fresh weights of Poa pratensis and P. tenuiflora were measured separately.

In the equation, Y A B represents the biomass of species A in mixed sowing, Y B A represents the biomass of species B in mixed sowing, Y A A represents the biomass of species A in monoculture, Y B B represents the biomass of species B in monoculture, Z A B represents the mixing proportion of species A in mixed sowing, and Z B A represents the mixing proportion of species B in mixed sowing (Fowler, 1982).

During the milk ripening stage of P. pratensis and P. tenuiflora in August 15, 2023, rhizosphere soil samples were collected for each treatment. Plants were carefully uprooted using a spade to ensure loose soil was shaken off and adhering soil was brushed away. The adhering soil was collected as rhizosphere soil. Subsequently, each rhizosphere soil sample was sieved through a 2 mm mesh to remove plant roots and other plant material. During sampling, sterile paper was used to wipe off residues adhering to the spade, and it was disinfected before collecting the next soil sample to prevent contamination between successive treatments and maintain sample freshness. The experiment had three replicates, with sampling conducted in an “S” shape, collecting 5 replicate samples from each plot, totaling 15 samples per treatment. A portion of each soil sample was air-dried and stored at room temperature for soil enzyme activity and soil nutrient analysis. Another portion of the samples was flash-frozen in liquid nitrogen, then transported from the field to the laboratory in a portable icebox and stored at -80°C for analysis of microbial diversity.

Soil Moisture Content (SMC) was assessed by drying the soil in an oven at a constant mass at 105°C. Soil Bulk Density (SBD) was calculated using the weight method (Al-Shammary et al., 2018). Total Nitrogen (STN) in the soil was determined using the Kjeldahl method (Chu et al., 2017), and Total Phosphorus (STP) was measured using the molybdenum-antimony anti-adsorption spectrophotometric method (CHEN et al., 2023). The measurement of Soil Urease (SU) activity involved incubating dry 10 g of soil with a 10% urea solution at 37°C for 24 hours. The formation of ammonium was determined spectrophotometrically at 578 nm. Soil Alkaline Phosphatase (SAP) activity was determined using the phenol disulfonic acid colorimetric method (Peng and Wang, 2016). Soil nitrate reductase (SNR) activity was measured using the phenanthroline-sulfuric acid colorimetric method (Hu et al., 2014).

Microbial DNA was extracted from 0.5 g of fresh soil four times (for a total of 2.0 g of soil) with a PowerSoil DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA, USA) following the manufacturer’s protocol. The purity and quality of the genomic DNA were checked on 0.8% agarose gels. The V3-V4 hypervariable region of the bacterial 16SrRNA gene was amplified with the primers 338 F (ACTCCTACGGGAG GCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT). For each soil sample, a 10-digit barcode sequence was added to the 5′ end of the forward and reverse primers (provided by Auwigene Company, Beijing). PCR was carried out on a Mastercycler Gradient Thermocycler (Eppendorf, Germany) using 50 μl reaction volumes containing 5 μl 10×Ex Taq Buffer (Mg2+ plus), 4 μl 12.5 mM dNTP Mix (each), 1.25 UEx Taq DNA polymerase, 2 μl template DNA, 200 nM barcoded primers 967 F and 1406R each, and 36.75 μl ddH2O. The cycling parameters were 94°C for 2 min, followed by 30 cycles of 94°C for 30 s, 57°C for 30 s and 72°C for 30 s, with a final extension at 72°C for 10 min. The fungal ITS region was amplified on an Eppendorf Mastercycler Gradient Thermocycler (Germany), with the primers ITS1F (5-CTTGGTCATTTAGAGGAAGTAA-3) and ITS2 (5-TGCGTTCTTCATCGATGC-3). The 5′ ends of both primers were tagged. The ultra-PAGE purified primers were ordered from Majorbio, China. The PCR mixtures were as follows: 4 μl 5× FastPfu Buffer, 1 μl each primer (5 μM), 2 μl dNTP mixture (2.5 mM), 2 μl template DNA, and 10 μlH2O. Thermocycling consisted of an initial denaturation at 95°C for 2 min, followed by 30 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s, with a final extraction at 72°C for 5 min. Three separate reactions of both bacterial and fungal samples were conducted to account for potentially heterogeneous amplification from the environmental template of each sample. PCR products were purified using the AXYGEN Gel Extraction Kit (QIAGEN) and quantified using qPCR. An equimolar mix of all three amplicon libraries was used for sequencing at Auwigene Company in Beijing, China.

The raw sequences of bacterial and fungal reads were initially trimmed using Mothur, and sequences that had the following three criteria were kept: (1) precise primers and bar-codes; (2) quality score>30; and (3) length>200 bp. The Ribosomal Database Project (RDP) classifi;er tool (Wang, 2007) was used to classify all sequences into different taxonomic groups. Qualified reads were separated using the sample-specific barcode sequences and trimmed with Illumina Analysis Pipeline Version 2.6. Then, the dataset was analyzed using QIIME. The sequences were clustered into operational taxonomic units (OTUs) at a similarity level of 97% to generate rarefaction curves (Colwell and Coddington, 1994) and to calculate the richness and diversity indices (Maidak et al., 1994).

Bacterial Functional Group Analysis: FAPROTAX is a database based on currently cultivable bacteria, containing over 7,600 functional annotations from multiple prokaryotes. It focuses more on predicting the biogeochemical cycles of samples. In this study, FAPROTAX (http://www.ehbio.com/ImageGP/index.php/Home/Index/FAPROTAX.html) was used for functional annotation of cultivable soil bacteria.

Fungal Functional Group Analysis: The soil fungal community was functionally annotated using the FUNGuild database (http://funguild.org). Confidence levels “probable” and “highly probable” were selected for the annotations.

Data analysis was conducted using SPSS 22.0 to assess the significance of differences in forage yield, soil nutrients, and soil enzyme activities among different planting years of cultivated grassland. A one-way analysis of variance (ANOVA) test was employed for this purpose. Figures were generated using Origin 9.1. Taxonomic alpha diversity was calculated as the estimated community diversity by the Shannon index using the Mothur software package (v.1.30.1). Nonmetric multidimensional scaling (NMDS) was selected to illustrate the clustering of different samples and further reflect the microbial community structure, while the changes in microbial structure under intercropping patterns were referred to as microbial beta diversity. Correlations among the soil microbial compositions, soil properties, and soil enzyme activities were determined using redundancy analysis (RDA). The RDA was performed using the CANOCO 4.5 software package. The relationships between the microbial characteristics (i.e., abundance, alpha diversity and beta diversity) and the soil properties and soil enzyme activities were determined by Spearman’s correlation analysis (SPSS 19.0, SPSS Inc., Chicago, USA). The data were analyzed using a one-way analysis of variation (ANOVA) for different intercropping patterns (P < 0.05), including soil properties, soil enzyme activities and the microbial characteristics (Wang Q. et al., 2023). The difference between mean values was determined using the least significant difference (LSD) (P > 0.05) as indicated by different letters. Origin 9.1 was used to draw figures.

The yield of forage in cultivated grassland is significantly influenced by the methods of sowing method and the duration of cultivation (Supplementary Table S1). Generally, the yield of forage in cultivated grassland shows a declining trend with the increase in years of cultivation. In the second year (2019) of cultivation, the yield of Puccinellia in monoculture grasslands was 152.07 g/m². As the cultivation period extended, the yield of Puccinellia in monoculture grasslands decreased by 40.10%, 13.81%, 25.68%, and 4.63% in the years 2020, 2021, 2022, and 2023, respectively. In the second year (2019) of cultivation, the yield of Poa in monoculture grasslands was 121.48 g/m². With the increase in cultivation years, the yield of Poa in monoculture grasslands decreased by 23.99%, 37.57%, 13.94%, and 4.63% in the years 2020, 2021, and 2022, respectively, but increased by 0.89% in 2023 (Supplementary Figure S1). In the second year of cultivation, the yield of mixed sowing grasslands was 123.40 g/m². With the increase in cultivation years, the yield of mixed sowing grasslands decreased by 12.21%, 13.16%, 20.21%, and 8.44% in the years 2020, 2021, 2022, and 2023, respectively. The experiments indicated that there was no significant change in the yield of forage in cultivated grassland from the fifth to the sixth year of cultivation (P<0.05). Moreover, the yield of Poa + Puccinellia in mixed sowing grasslands gradually exceeded the yield of Poa and Puccinellia with the increase in cultivation years. In the years 2019, 2021, 2022, and 2023, the yield of Poa + Puccinellia was higher than that of Puccinellia by 19.38%, 20.89%, 29.14%, and 23.97%, respectively. In 2019, the yield of Poa + Puccinellia was 18.86% lower than that of Puccinellia (Figure 1).

The interspecific relationships in mixed sowing grasslands exhibit variations. The relative biomass in mixed sowing cultivated grassland is predominantly located in the area above the RY_A = RY_B line, indicating that in these mixed grasslands, the growth of P. pratensis is inhibited (RY_A < RY_B). From 2019 to 2023, the relative yield of Poa + Puccinellia consistently lies between RT_A < 1.0 and RT_B > 1.0. This suggests that in mixed sowing grasslands, the intraspecific competition of P. pratensis is less than its interspecific competition (RY_A < 1.0), whereas for P. tenuiflora, the interspecific competition is less than its intraspecific competition (RY_B > 1.0). Consequently, P. tenuiflora has a competitive advantage over P. pratensis in mixed sowing grasslands. Furthermore, from 2021 to 2023, the relative yield of Poa + Puccinellia remains similar (Figure 2).

In 2019, the Relative Yield Total (RYT) of Poa + Puccinellia was significantly less than 1 (P<0.05), but with the increase in years of cultivation, the RYT of Poa + Puccinellia significantly exceeded 1 (P<0.05). This indicates that from 2021 to 2023, the level of competition in mixed sowing grasslands changed minimally, and the interspecific relationships within these grasslands stabilized (Figure 3). This suggests that in the second year of cultivation, there was intense competition in Poa + Puccinellia mixed sowing grasslands. However, as the cultivation period extended, the competition between P. pratensis and P. tenuiflora in these grasslands diminished. Both species, being perennial grasses and having similar plant heights, engaged in fierce competition for sunlight. Additionally, as both are grasses from the Poaceae family with similar nutrient requirements, they also competed intensely for soil nutrients. However, our experiment showed that with the increase in years of cultivation, the degree of interspecific competition gradually declined and tended towards stability. We hypothesize that this is due to the alteration of the soil microbial community when P. pratensis and P. tenuiflora are mixed sown, promoting nutrient cycling and soil enzyme activity in the soil. Therefore, we conducted tests on soil enzyme activity, soil nutrients, and soil microbial diversity in the Poa + Puccinellia mixed sowing grasslands in the sixth year of cultivation (Figure 4).

The soil properties of cultivated grassland are significantly influenced by the method of mixed sowing (Table 1; Supplementary Table S2). The soil moisture content (SMC), soil bulk density (SBD), and soil porosity (SP) are notably affected, with the SMC (P<0.05), SBD (P<0.05), and SP (P<0.05) showing significant variation. The SMC in mixed sowing grasslands of Poa + Puccinellia was higher by 34.65% compared to the Poa monoculture grasslands and 18.33% higher than the Puccinellia monoculture grasslands. The SBD of the Poa + Puccinellia mixed sowing grasslands was 6.34% lower than that of the Poa monoculture grasslands and showed no significant difference (P>0.05) when compared to the Puccinellia monoculture grasslands. The SP of Poa + Puccinellia was 5.26% higher than the Poa monoculture grasslands and 7.15% lower than the Puccinellia monoculture grasslands, although this difference was not significant (P>0.05) (Table 1).

Regarding soil nutrients, the soil total nitrogen (STN) in the mixed sowing grasslands of Poa + Puccinellia was 4.61% higher compared to the Poa monoculture grasslands and 7.27% higher than the Puccinellia monoculture grasslands. The STN in Poa + Puccinellia mixed sowing grasslands was 4.09% higher than in Puccinellia monoculture grasslands, but this difference was not statistically significant (P>0.05) when compared to the Poa monoculture grasslands (Supplementary Table S2; Figure 3).

In terms of soil enzyme activity, the soil nitrate reductase (SNR), soil acid phosphatase (SAP), and soil urease (SU) activities in Poa + Puccinellia mixed sowing grasslands were significantly higher than in Poa and Puccinellia monoculture grasslands. Specifically, the SNR in Poa + Puccinellia mixed sowing grasslands was 45.40% higher than in Puccinellia monoculture grasslands and 75.44% higher than in Poa monoculture grasslands. The SAP in Poa + Puccinellia mixed sowing grasslands was 10.91% higher compared to Puccinellia monoculture grasslands and 9.65% higher than Poa monoculture grasslands. Similarly, the SU in Poa + Puccinellia mixed sowing grasslands was 13.79% higher than in Puccinellia monoculture grasslands and 11.61% higher than in Poa monoculture grasslands (Supplementary Table S2; Figure 5).

The Goods Coverage indices for bacterial and fungal communities in the soils of Puccinellia, Poa + Puccinellia, and Poa grasslands all reached 1. This indicates that the sequencing results comprehensively represent the true state of microbial diversity in the rhizosphere of these grasslands. Experimental results show that compared to monoculture Puccinellia and Poa grasslands, the Poa + Puccinellia mixed grassland has no significant effect on the Chao1 index, ACE, sobs, and Pielou index of bacterial communities (P>0.05) (see Supplementary Table S3), but has significant effects on the Shannon and Simpson indices (Table 2). This suggests that the Poa + Puccinellia mixed grassland has a minimal impact on the abundance and evenness of bacterial communities but significantly increases the richness of bacterial communities (P<0.05). For fungal communities, the Poa + Puccinellia mixed grassland reduces the Chao1 index and ACE but has no significant effect on sobs, while significantly increasing the Pielou and Shannon indices (P<0.05). This indicates that the Poa + Puccinellia mixed grassland has a minimal impact on the abundance of fungal communities but significantly enhances the evenness and richness of fungal communities (P<0.05) (Table 2).

Principal Coordinates Analysis (PCoA) and Analysis of Similarities (ANOSIM) were employed to assess the beta diversity of microbial communities (Figures 6, 7). Principal coordinate 1 accounted for 13.24% and 18.95% of the total variation in the two groups, respectively, while principal coordinate 2 accounted for 12.85% and 17.23%. Moreover, ANOSIM revealed significant differences in the microbial abundance of soil bacteria and fungi among Puccinellia, Poa + Puccinellia, and Poa grasslands (R=0.5638, P=0.005; R=0.786, P=0.003) (Figures 6, 7).

In all grasslands, the dominant bacterial phyla with abundance greater than 1% were consistent, including Proteobacteria, Acidobacteriota, Planctomycetota, Bacteroidota, Actinobacteriota, Gemmatimonadota, Chloroflexi, Verrucomicrobiota, Patescibacteria, and Myxococcota (Figure 8). Similarly, the dominant fungal phyla with abundance greater than 1% were Ascomycota, Mortierellomycota, Basidiomycota, and Olpidiomycota (Figure 8; Supplementary Table S4). The mixed sowing of Poa + Puccinellia significantly affected the abundance of bacterial phyla such as Proteobacteria, Acidobacteriota, Bacteroidota, Gemmatimonadota, Verrucomicrobiota, and Myxococcota compared to Puccinellia monoculture grasslands (P<0.05), but the effect was not significant compared to Poa monoculture grasslands (P>0.05). However, compared to Poa monoculture grasslands, the mixed sowing of Poa + Puccinellia significantly affected the abundance of fungal phyla such as Ascomycota, Mortierellomycota, Basidiomycota, and Olpidiomycota (P<0.05), while the effect was not significant compared to Puccinellia monoculture grasslands (P>0.05) (Figure 9; Supplementary Table S4).

Functional Annotation of Prokaryotic Taxa (FAPROTAX) was used to annotate the functional capabilities of bacteria in cultivated grassland. The results indicated that mixed sowing grasslands increased he gene abundance for nitrogen fixation and aromatic compound degradation. However, there was a reduction in gene abundance for aromatic hydrocarbon degradation, aliphatic non-methane hydrocarbon degradation, and hydrocarbon degradation overall (Figure 10).

Furthermore, FunGuild was employed for the functional annotation of fungi in cultivated grassland. The results showed that mixed sowing grasslands increased the gene abundance for various functional groups, including Animal Pathogen-Endophyte-Lichen Parasite-Plant Pathogen-Soil Saprotroph-Wood Saprotroph, Endophyte-Litter Saprotroph-Soil Saprotroph-Undefined Saprotroph, Animal Pathogen-Dung Saprotroph-Endophyte-Lichen Parasite-Plant Pathogen-Undefined Saprotroph, Endophyte-Plant Pathogen-Wood Saprotroph, Endophyte-Epiphyte-Fungal Parasite-Insect Parasite, and Fungal Parasite-Undefined Saprotroph. Conversely, there was a decrease in the gene abundance for Plant Pathogen, Plant Saprotroph-Wood Saprotroph, Fungal Parasite-Plant Pathogen-Plant Saprotroph, Animal Pathogen-Endophyte-Plant Pathogen-Wood Saprotroph, Animal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood Saprotroph, and Animal Pathogen-Plant Pathogen-Undefined Saprotroph (Figure 10).

Redundancy analysis (RDA) demonstrated that soil nutrients and enzyme activities have different responses to variations in bacterial and fungal phyla. There is a significant positive correlation between soil total nitrogen (STN) and soil total phosphorus (STP), as well as soil enzyme activities such as soil nitrate reductase (SNR), soil acid phosphatase (SAP), and soil urease (SU) with the variation in bacterial Acidobacteriota, Verrucomicrobiota, and the fungal phylum Mortierellomycota (Figure 11).

Pearson’s correlation analysis indicates that STN, SU, SNR, SAP are significantly correlated with the yield of cultivated grassland (P<0.05). Using the Mantel test, it was found that SMC, SP, SNR, SAP are significantly correlated with bacterial communities, and yield, SMC, SNR, SAP, SU, STN are significantly correlated with fungal communities (Figure 12).

Previous studies have shown that high seeding density in Cultivated grasslands leads to intense competition and plant mortality. However, as cultivation time increases and plant density decreases, competition among plants in cultivated grasslands stabilizes (Schippers and Kropff, 2001; Partzsch, 2011). Our research demonstrates that the Relative Yield of P. tenuiflora and P. pratensis in mixed sowing grasslands tends to stabilize with increasing cultivation years. Furthermore, the forage yield of all treatments shows a decreasing trend over time, but stabilizes from 2021 (the 4th year of cultivation) to 2023 (the 6th year of cultivation). These findings are consistent with previous research, suggesting that with increasing cultivation time, all treatments in cultivated grasslands reach a stable competitive state by 2021 (the 4th year of cultivation). Furthermore, our experiment indicates that in 2021 (the 4th year of cultivation), the forage yield in cultivated grasslands stabilizes. Additionally, the forage yield of mixed sowing grasslands of P. tenuiflora and P. pratensis is significantly higher than that of monoculture grasslands of P. tenuiflora and P. pratensis. Therefore, mixed sowing of P. tenuiflora and P. pratensis can increase the forage yield of grasslands.

Numerous studies have demonstrated that greater niche overlap leads to more intense interspecific competition for limited resources, resulting in reduced community stability and lower yields (Baillard et al., 2021; Reiss and Drinkwater, 2022; Glowacki et al., 2023). Our study findings reveal that in 2019 (the second year of cultivation), the forage yield of Poa + Puccinellia mixed sowing grasslands was lower than that of monoculture grasslands, with a significantly lower Relative Yield Total (RYT<1.0, P<0.05). This could be attributed to the excessive planting density during the early growth stages and the intense competition for photosynthesis and nutrient uptake between P. tenuiflora and P. pratensis, both of which belong to the Poaceae family and are classified as bottom leaf grasses. These grasses have similar plant heights and nutrient requirements, resulting in fierce competition for light and nutrients (Golińska et al., 2023; Skuodienė et al., 2023). However, as the mixed sowing grasslands grew, the RYT increased to greater than 1.0 (P<0.05), indicating an increase in interspecific compatibility and a decrease in competition. This change can be attributed to the growth of species in the mixed sowing grassland over time, a decrease in planting density, and an increase in available ecological factors, resulting in a reduction in ecological niche overlap between competitors and a reduction in interspecies competition (Niu et al., 2009). Therefore, the forage yield in the mixed sowing grassland gradually surpasses that of monoculture grasslands of P. tenuiflora and P. pratensis.

Previous studies have shown that when the activity of soil alkaline phosphatase is enhanced, the rate of conversion of organic phosphorus to inorganic phosphorus in the soil accelerates, increasing the content of available phosphorus (Kuzmanović et al., 2024; Rocabruna et al., 2024). An increase in the activity of soil nitrate reductase can affect the rate of soil denitrification and the amount of nitrogen available to plants (Li et al., 2024; Paludo et al., 2024). Soil urease catalyzes the hydrolysis of urea into NH₃, which is further converted to NH⁴⁺ through proton exchange (Qin et al., 2010; Pinto et al., 2014; Chen et al., 2021; Duff et al., 2022). Consistent with previous studies, our experiment demonstrated a significant positive correlation between the total nitrogen content of mixed sowing grasslands and the activities of nitrate reductase and urease. Additionally, we observed a significant positive correlation between the total phosphorus content and the activity of soil alkaline phosphatase. This indicates that the mixed sowing grassland of P. tenuiflora and P. pratensis can promote soil nutrient cycling by increasing the enzymatic activity of soil alkaline phosphatase, soil nitrate reductase, and soil urease.

A plethora of studies have consistently demonstrated that mixed sowing grasslands can enhance soil nutrient availability by decomposing plant litter into mineralized nutrients, thereby increasing the nitrogen and phosphorus content in the soil (Horrocks et al., 2016; Husse et al., 2017; Sollenberger et al., 2019). Additionally, the secretion of organic acids and enzymes by the root system of mixed sowing grasslands can facilitate the release of locked nitrogen and phosphorus in the soil, ultimately improving their availability (Li et al., 2023; Wang Z. et al., 2023). Moreover, mixed sowing grasslands can contribute to the enhancement of soil nitrogen-fixing capabilities by improving the soil microbial community (Philippot et al., 2023). These pathways collectively contribute to the increase in soil total nitrogen and total phosphorus content. Consistent with prior research findings, the soil total nitrogen and total phosphorus content in P. tenuiflora and P. pratensis mixed sowing grasslands were significantly higher compared to monoculture grasslands. This suggests that cultivated grasslands established through the mixed sowing of P. tenuiflora and P. pratensis can effectively enhance the total nitrogen and total phosphorus content in the soil.

However, previous studies have indicated that increasing or altering the ecological factors utilized by species can reduce niche overlap among competing species, thereby mitigating interspecific competition and benefiting the stability of competitive communities (Morales et al., 2023). Our study shows that the levels of soil total nitrogen (STN) and soil total phosphorus (STP) in mixed sowing grasslands are significantly higher than in monoculture grasslands of P. tenuiflora and P. pratensis (P<0.05). By 2023 (the sixth year of cultivation), the forage yield of mixed sowing grasslands had surpassed that of the monoculture P. tenuiflora and P. pratensis grasslands, with a Relative Yield Total (RYT) greater than 1.0 (P<0.05). This can be attributed to the higher levels of STN and STP in the soil of Poa + Puccinellia mixed sowing grasslands compared to monoculture grasslands. These elevated nutrient levels help promote cell division and leaf expansion, increasing leaf area, beneficial for capturing light energy and enhancing photosynthesis. Additionally, it enables plants to develop more extensive root systems, advantageous for maximizing the absorption of soil moisture and minerals, leading to niche differentiation between the two plants and reducing interspecific competition in the mixed sowing grasslands.

The microbial diversity index, a key indicator of species richness in soil microbial communities, suggests that higher indices correlate with richer microbial diversity, more complex ecosystems, and greater functional stability (Deng, 2012; Bastida et al., 2021; Osburn et al., 2023). Our study shows that mixed sowing significantly increases soil bacterial alpha diversity (P<0.05). indicating that mixed sowing enhances bacterial diversity, ecosystem complexity, and functional stability compared to monoculture grasslands (Diakhaté et al., 2016).

Our investigation reveals that mixed sowing significantly modulates the functional gene composition in soil bacterial communities, notably enhancing the frequency of genes implicated in the degradation of aromatic compounds and in nitrogen fixation processes. Aromatic compounds are ubiquitously present in nature, including in all biological entities, and their degradation predominantly occurs through microbial processes that transform complex organic matter into forms utilizable by microbes and plants (Shen et al., 2012; Shahsavari et al., 2019; Singh and Kumar, 2019). The increased abundance of genes associated with aromatic compound degradation in the bacterial communities of mixed sowing grasslands indicates that mixed sowing enhances the activity of soil bacteria in decomposing aromatic compounds, thereby improving soil fertility and productivity (Li et al., 2020; Wang Z. et al., 2023). Furthermore, the increased abundance of soil nitrogen fixation genes can enhance the soil’s nitrogen fixation capacity, increase soil nitrogen levels, and thus alleviate nutrient competition in mixed sowing grasslands (Dai et al., 2018).

Our research findings indicate that mixed sowing grasslands significantly decrease the abundance of the Ascomycota phylum within the fungal community and considerably increase the abundance of the Mortierellales phylum. The Ascomycota phylum is a critical driver of carbon and nitrogen cycling in arid ecosystems (Wang Z. et al., 2022). Previous studies have shown that long-term application of nitrogen fertilizers promotes the growth of copiotrophic microorganisms in the soil while inhibiting oligotrophic microorganisms (Challacombe et al., 2019). This aligns with our study’s results, where mixed sowing of Poa + Puccinellia grasslands increases soil total nitrogen content, leading to a reduction in the abundance of the Ascomycota phylum.

The Mortierellales phylum plays a significant role in soil nutrient cycling (Davies et al., 2010). Our study indicates a positive correlation between the Mortierellales phylum and soil STN, SAP, SU, STP, aligning with prior research findings. Additionally, through the Mantel test, we demonstrated a significant positive correlation between yield in mixed sowing grasslands and soil fungal community structure. Therefore, our study suggests that mixed sowing can enhance soil nutrient transformation efficiency by improving the soil fungal community, thereby increasing soil nutrients and reducing interspecific competition in mixed sowing grasslands.