Exchangeable cations determined by different methods: Implications for soil classification and evaluation

For a comprehensive and global evaluation of soils, it is essential to define common diagnostic properties among classification or interpretative systems. Specifically, in soil taxonomy, the Brazilian Soil Classification System (SiBCS) adopts analytical methods that differ from those used in international systems such as the World Reference Base for Soil Resources (WRB) and the Soil Taxonomy (ST). In the chemical properties, differences are observed in the determination of exchangeable basic cations and potential acidity (H+Al), as well as derived properties such as the sum of bases (SB), cation exchange capacity (CEC pH7 ), and base saturation (V). This study aimed to: (i) to compare the values of H+Al, SB, CEC, and V in Brazilian soils determined by different analytical methods; (ii) to propose regression models to harmonize values of the properties obtained through different analytical procedures; and (iii) to compare the threshold values of diagnostic properties used as classification criteria in the SiBCS with those adopted by WRB and ST. A representative database of Brazilian soils published in the Brazilian Soil Data Repository (SoilData) was used, comprising 2,217 samples from surface horizons and 3,726 from subsurface horizons. Values of SB, CEC, and V for surface and subsurface horizons, obtained using the methods applied in the SiBCS, were lower than those determined by the WRB and ST methods. The difference in the estimated V values was below 5 %, which falls within the analytical error range. Thus, the SiBCS criteria of V ≥50 % for identifying eutrophic character and >65 % for defining a chernozemic A horizon were found to be high, and they tend to underestimate several Brazilian soil classes, particularly those classified as Chernossolos, as well as the misclassification of anthropic and other soils modified by agricultural practices. It is proposed that the SiBCS adopt thresholds similar to those used in WRB and ST for base saturation, or alternatively, apply regression equations to improve correspondence among the systems.