3.1.1 Characteristics of soil microbial carbon and nitrogen in different H. rhamnoides forests
As shown in Table 2, our analysis revealed that soil microbial C and N content in different soil layers showed a trend of small change with increasing soil layers. That is, soil microbial activities weakened gradually as the soil depth increased. There was a slight increase in soil microbial C content in the 60–80 cm soil layer of H. rhamnoides pure forest (shady slope), the 80–100cm soil layer ofH. rhamnoides pure forest (sunny slope), the 80–100 cm soil layer of H. rhamnoides and Chinese arborvitae mixed forest and the 60–80 cm soil layer of H. rhamnoides and Pinus tabulaeformis mixed forest. There was no clear vertical variation of soil microbial N content, and there was no significant difference between different soil layers. Over the four plots, the soil microbial C content was highest in H. rhamnoides pure forest (sunny slope), whereas there was little difference in the soil microbial C contents of the other three plots. Soil microbial N content was highest in H. rhamnoides pure forest (sunny slope), and the lowest value was inH. rhamnoides and P. tabulaeformis mixed forest.
3.1.2 Regression analysis of soil microbial carbon and nitrogen contents and soil carbon and nitrogen contents in different H. rhamnoides forests
Regression analysis, as a method for statistical analysis of the inter-dependent quantitative relationship between two or more variables, can be used to analyze the relationship between one variable and the change of another variable. We analyzed the soil microbial C and N content and the soil C and N content in different H. rhamnoidesforests via exponential and linear regression analysis and established a regression model. Table 3 shows the results. Apart from the H. rhamnoides and P. tabulaeformis mixed forest, the regression equations of the other three samples were well fitted. Of the three well-fitting plots, the R2 of the index model of soil microbial biomass N and alkali-hydrolysable N of H. rhamnoidespure forest (sunny slope) was 0.73, whereas the remainder were all above 0.8, with some even exceeding 0.9. Therefore, the model can well explain the relationship between soil microbial C content and soil organic C content, soil microbial N content and soil total N content, and soil microbial N content and alkali-hydrolyzed N content. Only the soil microbial N content and soil alkali-hydrolyzed N content regression model showed a good fit in the case of H. rhamnoides and P. tabulaeformis mixed forest.