Strengths of relationships among soil microbial and organic matter properties are scale-dependent
Simon, E.; Guseva, K.; Alteio, L. V.; Kaiser, C.
Show abstract
Relationships among variables in ecological systems are inherently scale-dependent, especially in heterogeneous systems. Yet it remains to be examined whether relationships among variables vary across observation scales in soil. Generally, it is desirable that observation scale matches the intrinsic scale of a process or pattern. Millimetre-sized soil aggregates are closer to the intrinsic scale of microbial communities than traditionally studied bulk soil samples, making them more suitable for studying potential links between microbial communities and their environment. To explore the effect of observation scale on relationships among soil parameters, we measured bacterial, archaeal, and fungal taxa richness and density, organic matter properties (e.g., carbon and nitrogen content, stoichiometric and isotopic ratios), and soil water content in individual aggregates and aliquots of homogenised soil cores, bulk soil samples, in two soil layers. We analysed pairwise correlations among these variables and assessed whether individual aggregates systematically differed from bulk soil samples. Organic matter properties were more strongly correlated in bulk soil samples, consistent with the idea that increasing the sample volume reduces noise. In contrast, microbial community and organic matter properties showed weaker correlations in bulk soil samples than aggregates in topsoil. In addition, we found that aggregates and bulk soil samples differed systematically in individual microbial and organic matter properties, particularly in the topsoil. Our study demonstrates that relationships among variables in soil are spatial scale-dependent. Aggregates offer valuable insights into microbial communities in soil, complementing bulk soil samples, and are useful for studying links between microbial communities and their environment.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.