Use of Bayesian Methods to Model Soil Aggregation in Undisturbed Semiarid Grasslands

Comparisons of soil aggregation and aggregate‐protected organic matter among soil textures in the same climatic zone and of the same mineralogical composition are somewhat limited. Two undisturbed, semiarid sites in northeastern Wyoming were selected to determine the differences in aggregation and organic matter between loam and sandy clay loam soils. Additionally, data from these two sites were used to illustrate differences between two statistical approaches to data analysis: Bayesian modeling in the context of a two‐sample comparison of means and conventional t ‐tests. Examination of posterior distributions (Bayesian modeling) and confidence intervals (conventional t ‐tests) resulted in similar conclusions about the influence of soil texture on aggregation and C and N concentrations. Loam soils generally had more structure (as indicated by higher macroaggregation) and aggregate‐protected C and N than sandy clay loam soils, which had greater amounts of available C and N not protected by aggregates. For soils in undisturbed semiarid grasslands with similar clay contents, in the range of 11 to 24%, we suspect that aggregate stability was primarily affected by sand content, while aggregate C and N accumulations were influenced by silt. A higher percentage of silt in the loam soils appeared to have provided enough of a difference in small pore space between the two soil types to result in higher C and N storage in loam soils relative to sandy clay loam soils. Soil texture is a major driving force behind aggregation as well as C and N storage in semiarid soils of the same mineralogy.