土壤中微生物含量影响因素的统计方法分析

Statistical analysis of influencing factors of soil microbial content

摘要:

【目的】采用多元方差分析、多元线性回归分析、通径分析等统计方法，并利用多元统计方法对土壤中微生物含量进行分析。【方法】通过研究这3种常用统计方法的原理、应用条件、适用于解决的问题等，用不同的方法分析影响土壤中微生物含量的各种因素，利用SPSS软件求解。【结果】得出不同降雨量下土壤中的微生物含量存在差异性；铵态氮含量和海拔高度与变形菌门含量呈显著的线性关系，土壤中变形菌门含量随着铵态氮含量增多而增多，且铵态氮含量对变形菌门含量所起的直接作用最大；土壤中铵态氮含量、总氮量和海拔高度与放线菌门含量呈显著的线性关系，且均为负相关；其中总氮量对土壤中放线菌门含量的变化所起的负面作用最大，但铵态氮含量对放线菌门含量所起的直接作用最大。对比描述性统计方法与多元方差分析方法可知，应用描述性统计方法只能得出降雨量是影响土壤中不同细菌群落相对含量的因素之一，而不知降雨量对微生物含量变化的影响程度，6个梯度降雨量引起地显著差异究竟是由哪种微生物引起的，方差分析方法可以解决以上的问题。对比多元线性回归分析方法与通径分析方法可知，多元线性回归分析得出的相关系数只能表示铵态氮含量、总氮量和海拔与变形菌门含量(放线菌门含量)之间关系的密切程度，但无法解释和分析这种关系的构成和来源，通径分析方法可以解决这个问题。【结论】通径分析方法在分析相关问题时要优于多元线性回归分析方法，通径分析比多元线性回归分析方法更能体现因素对变形菌门含量(放线菌门含量)的直接影响和间接影响，分析结论更直观、更能说明问题。

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.