Statistical analysis of influencing factors of soil microbial content

Abstract:

[Objective] We evaluated statistical methods of multivariate variance, multiple linear regression and path analysis and used multivariate statistical methods to analyze the microbial content in soil. [Methods] Through researching principle, application conditions and application to solve the problem of these three common statistical methods, we used different methods to analyze the impact of various factors on the microbial content in soil, and used SPSS software to solve. [Results] We found that there were differences in the amount of microbial content in the soil under different rainfall. Ammonium nitrogen content and altitude with proteobacteria content showed linear relationship. In soil, proteobacteria content increased with the increase of ammonium nitrogen content, and the direct effect of ammonium nitrogen content on proteobacteria content was most important. Ammonium nitrogen content, total nitrogen, altitude with actinobacteria content in soil showed adverse linear relationship, and the negative effect of total nitrogen content on actinobacteria content in soil was the most, but the direct effect of ammonium nitrogen content on actinobacteria content was the most. Compared with descriptive statistics and multivariate variance analysis showed that using descriptive statistical methods only draw rainfall was one of the factors of the relative content of different bacterial communities in soil and didn’t know the impact of rainfall on the changes of content of microorganisms and what the cause of the significant difference between the 6 gradients of rainfall. Variance analysis methods could solve the above problems. Compared with multiple linear regression analysis and path analysis showed that correlation coefficient obtained by multiple linear regression analysis only said ammonium nitrogen content, total nitrogen content and altitude and deformation of bacterial phyla content (actinobacteria content) the relationship between the close degree, but unable to explain and analyze the composition and source of the relationship, path analysis methods could solve this problem. [Conclusion] Path analysis method in the analysis of relevant problems was superior to multiple linear regression analysis method. Path analysis more than multiple linear regression analysis methods could reflect the direct and indirect impact of factors on the proteobacteria content (actinobacteria content). Its analysis conclusion was more intuitive, able to explain the problem.