Building a Quantitative Analogy for Soil Classification Systems.

Soil heterogeneity is a major contributor to the uncertainty in environmental fate modeling. Even if the modeling is executed carefully with full experimental validation, predictions will apply only to a particular soil type(s) associated with the specific site of interest, making transferability to other sites questionable. With this research, we sought to overcome this limitation by developing a new mathematical analogy that can be attached to a soil classification system. The idea is to “transcribe” the inherent but largely qualitative criteria in the taxonomic system to quantifiable physicochemical descriptions. Based on the NRCS soil classification system, we collected soil horizons classified under the Alfisols taxonomic Order. These samples were described via extensive physical and chemical characterizations. Using the multivariate statistical modeling, we developed a quantitatively descriptive analogy from the characterization data, and tested its ability to statistically discriminate the samples at the Suborder and Great Group taxonomic sublevels. Our results showed strong latent structure in the soil physicochemical characterization data, with major soil variables contributing to that structure included exchangeable cations, cation exchange capacity, soil pH, solid-phase carbon, nitrogen, and sulfur, and texture. The analogy successfully discriminated 84% of the soils among three different Alfisols Suborders: Udalfs, Aqualfs, and Ustalfs. At the more specific Great Group sublevel, the analogy successfully classified 63% of the soils, with ambiguity arising from the poorly defined, “catch-all” Hapludalfs sublevel, thus identifying opportunities to meaningfully refine this designation based on geochemical data. Assuming such latent structure can be articulated throughout the NRCS taxonomic system, these results demonstrate for the first time the potential for developing quantitative analogical models predicting complex soil biogeochemical behavior for a vast array of different soil types.