Soil nutrient contents
As shown in supplementary Table S1, the water content and pH of the living cover was considerably (p < 0.05) higher than the no cover treatment. In the upper soil samples, the water content of living cover treatment was 11.59% (17.23 ± 0.36%; 15.44 ± 0.36%) and 19.43% (22.50 ± 1.75%, 17.45 ± 0.95%) higher than that of no cover treatment in May and July respectively. While in the deeper soil, they are 28.94% (19.36 ± 0.98%, 16.21 ± 0.69%) and 48.05% (28.78 ± 0.98%, 19.44 ± 0.58%), respectively. In May, the pH of living cover treatment was 14.53% (6.07 ± 0.06, 5.30 ± 0.10) and 9.88% (6.23 ± 0.06, 5.67 ± 0.12) higher than that of no cover treatment in the upper soil and deeper soil respectively. And in July, it was 6.37% (6.17 ± 0.06, 5.80 ± 0.10) and 10.70% (6.83 ± 0.06, 6.17 ± 0.06) respectively. The soil TOC content under living cover treatment was higher than before (11.37 g/kg), but had no significant (p > 0.05) effect under no cover treatment. Compared with no cover treatment, the TOC content of living cover treatment in the upper soil increased by 18.02% (14.08 ± 0.59 g/kg, 11.93 ± 0.74 g/kg) and 17.21% (13.35 ± 0.51 g/kg, 11.39 ± 0.89 g/kg) respectively in May and July, while that in the deeper soil increased by 32.87% (12.37 ± 0.33 g/kg, 9.31 ± 0.88 g/kg) and 14.82% (10.15 ± 0.39 g/kg, 8.84 ± 0.30 g/kg) respectively. In May, the levels of TP, TN, AP and AK increased by 13.20%, 8.89%, 13.84% and 13.79%, respectively, while in July, they were 12.17%, 12.12%, 16.82% and 23.10%. And with the change of seasons, TN content decreased in two soil layers, but TP, AP and AK contents increased.
Soil Enzymes
The living cover treatment significantly (p < 0.05) improved the soil urease (URE), catalase (CAT), alkaline phosphatase (ALP), invertase (INV) activity contents (Additional file 2: Table S2). In the upper soil, the soil urease activity of living cover treatment (11.60 ± 0.13 IU/g) was 22.88% more than that of no cover treatment (9.44 ± 0.07 IU/g) in May and 49.15% in July (12.35 ± 0.20 IU/g, 8.28 ± 0.04 IU/g). And the soil catalase activity of living cover treatment was 12.28% (187.50 ± 2.78 IU/g, 167.00 ± 2.02 IU/g) higher than that of no cover treatment soil in May and 33.38% (194.48 ± 2.84 IU/g, 145.81 ± 2.35 IU/g) in July. And the soil alkaline phosphatase activity of living cover treatment was 5.33% (0.79 ± 0.01 IU/g, 0.75 ± 0.02 IU/g) higher than that of no cover treatment soil in May and 11.76% (0.76 ± 0.01 IU/g; 0.68 ± 0.02 IU/g) in July. The soil invertase activity of living cover treatment was 50.94% (0.80 ± 0.03 IU/g, 0.53 ± 0.02 IU/g) higher than that of no cover treatment soil in May and 21.43% (0.85 ± 0.04 IU/g, 0.70 ± 0.01 IU/g) in July. In the deeper soil, the soil urease activity of living cover treatment was 1.05 times (8.37 ± 0.10 IU/g, 7.90 ± 0.17 IU/g) that of no cover treatment soil in May and 1.10 times (8.11 ± 0.15 IU/g, 7.36 ± 0.17 IU/g) in July. And the soil catalase activity of living cover treatment was 1.30 times (166.87 ± 1.28 IU/g, 128.33 ± 3.84 IU/g) that of no cover treatment soil in May and 1.11 times (151.76 ± 1.38, 136.66 ± 1.50 IU/g) in July. The soil alkaline phosphatase activity of living cover treatment was 1.22 times (0.73 ± 0.01 IU/g, 0.60 ± 0.02 IU/g) that of no cover treatment soil in May and 1.17 times (0.56 ± 0.03 IU/g, 0.48 ± 0.02 IU/g) in July. The soil invertase activity of living cover treatment was 1.52 times (0.70 ± 0.02 IU/g, 0.46 ± 0.03 IU/g) that of no cover treatment soil in May and 1.13 times (0.72 ± 0.02 IU/g, 0.64 ± 0.01 IU/g) in July. The enzyme activities of each treatments were higher in the upper soil than that in the deeper soil.
Fungal Community Diversity
The Shannon index (H'), Richness index (S) and Evenness index (E') of fungal community were evaluated with the Illumina Miseq high-throughout sequencing data (Additional file 3: Table S3). Most living cover treatments remarkably increased the levels of H', S, and E' in the soil. And living cover treatment can significantly (p < 0.05) increase the richness index (S). Living treatment (189.7 ± 6.4, 188.7 ± 18.9) was 38.91% higher than no cover treatment (136.7 ± 8.5, 135.7 ± 16.2) in spring while it was 38.80% in summer (171.0 ± 8.5, 161.7 ± 13.4; 115.0 ± 5.6, 124.7 ± 10.8). Among the no cover treatments and living cover treatments, there was no remarkably (p > 0.05) difference in the three diversity indexes between the upper and deeper soil. And the season had no obvious influence on the three indexes. The S level of living cover treatment was the highest in the upper soil in May (189.7 ± 6.4), which was considerably (p < 0.05) higher than the no cover treatments.
Bacterial Community Diversity
The Shannon index (H'), Richness index (S) and Evenness index (E') of bacterial community were evaluated by the Illumina Miseq high-throughout sequencing data (Additional file 4: Table S4). All living cover treatments improved the H', S, and E' in the soil but there was no obvious (p > 0.05) difference in these three diversity indexes between the no cover treatments and living cover treatments expect NB_May (392.7 ± 24.8) and LB_May (440.7 ± 33.8). And the season and the soil depth have no significant (p > 0.05) influence on the three indexes. The S level of living cover treatment (440.7 ± 33.8) was the highest in the deeper soil in May.
Fungal Community Structure
To standardize the different sequencing depths, each sample made random selection of 44,412 reads to analysis (Additional file 5: Table S5). Ascomycota (60.16%) and Basidiomycota (25.36%) were the main fungal communities (Fig. 1). Unclassifled_k_Fungi was represented in 3.14% of all sequences. The relative abundances of Ascomycota and Mortierellomycota of the living cover treatments decreased compared with that in no cover treatments, but relative abundances of Basidiomycota increased. Principal component analysis (PCA) also showed the difference between no cover and living cover treatments, and the soil samples in different treatments were obviously separated. Two principal components were determined, which clarified the total variance of 99.29% in the dataset (Fig. 2). Sequences from the Ascomycota dominated in upper soil samples (63.31%) and deeper soil samples (57.01%) collected from both two treatments (no cover and living cover) (Fig. 1a). With the change of seasons, the relative abundances Ascomycota decreased, whereas that of the Basidiomycota increased under two treatments (no cover and living cover) in phylum level, but there was no obvious change in Mortierellomycota (Fig. 1a). The relative abundance of Ascomycota in spring was higher than that in summer, but the relative abundance of Basidiomycota was opposite.
The overall fungal community was dominated by Sordariomycetes (40.92%) and Agaricomycetes (14.43%) (Fig. 1b) under the class level. In the both all treatments, the class Agaricomycetes and Tremellomycetes were represented by 18.89% and 18.95% of the sequences in living cover sampling, whereas the class represented only 9.98% and 2.70% in no cover (Fig. 1b). And the the class Mortierellomycetes and Eurotiomycetes were represented by 15.93% and 10.07% of the sequences in no cover sampling, whereas the class represented only 4.96% and 4.02% in living cover (Fig. 1b). With the seasons change, the relative abundances of Sordariomycetes of the living cover treatments decreased while in deeper soil increased. The relative abundances of Agaricomycetes, Tremellomycetes and Leotiomycetes of the living cover treatments increased compared with that in no cover treatments in both upper soil and deeper soil, while the situation of Mortierellomycetes, Eurotiomycetes and Pezizomycetes were the opposite (Fig. 1b).
Bacterial Community Structure
To standardize the different sequencing depths, each sample made random selection of 14,275 reads to analysis (Additional file 6: Table S6). The dominant fungi are Proteobacteria (44.78%), Actinobacteria (16.48%) and Acidobacteria (15.22%) among all samples (Fig. 3). The living cover was not obviously affected the relative abundances of the phylum expect Elusimicrobia (Additional file 7: Figure S1). The relative abundances of Acidobacteria of the living cover treatments decreased compared with that in no cover treatments while relative abundances of Actinobacteria increased but all of these two phylum were not demonstrated a significant correlation. And there was no significant (p > 0.05) difference between deeper soil and upper soil while there were significant differences between seasons in phylum and class level (Additional file 8,9: Figure S2, Figure S3).
The result of PCA showed the obvious difference between soil samples in spring and summer. And two principal components were determined, which can explain 88.38% of the total variance in the dataset (Fig. 4). Sequences from Proteobacteria dominated in Spring soil samples (37.79%) and Summer soil samples (51.77%) collected from both two seasons (Fig. 3). The relative abundance of Proteobacteria and Acidobacteria were the highest in Summer, and the lowest in Spring while Actinobacteria and Firmicutes reached its highest abundance in Spring.
Relationship Between Microbial Community Structures And Soil Properties
Living cover treatment changed microbial community structures and soil properties. After removal of the redundant variables, ten soil properties were chosen for RDA. As shown in Fig. 5a,b, pH, TOC, TP, TN, AK, URE and INV significantly affected the fungal community structure in upper soil, while soil properties except AK and URE all obviously (p < 0.05) impacted the fungal community composition in the deeper soil. Soil properties except TOC, TN and ALP all remarkably (p < 0.05) impacted the bacterial community composition in the upper soil, while SWC, pH, AP and ALP dramatically (p < 0.05) impacted the bacterial community composition in the deeper soil (The relevant p value was in Additional file 10: Table S7). From the angle between arrow connecting lines representing different soil properties, TP, pH, INV, AK, AP, and SWC were always small, which showed that they had a good correlation and were positively correlated in each treatment. In the same way, TN, TOC, ALP, and URE were also positively correlated.
Prediction Of Community Functions Of Soil Fungi And Bacteria
The micro-ecological functions of fungi and bacteria in the soil of hazelnut orchard under no cover and living cover treatments were studied by analyzing the fungal and bacterial communities by FUNGuild and PICRUSt1. The guilds identified in present study were listed in Fig. 6, the consequences demonstrated that the functional prediction results of fungi were related to different treatments. And three nutritional patterns (pathotrophs, saprotrophs and symbiotrophs) accounted for approximately 36.69%, 49.80% and 2.59% of no cover treatment fungal OTUs respectively whereas in living cover treatment was 19.76%, 41.51% and 6.89%. There was no obvious divergence between no cover treatment and living cover treatment according to analyzing the bacterial communities in the Cluster of Orthologous Groups (COG) database by PICRUSt1 (Additional file 11: Figure S4).