In a salt-affected alluvial plot, increased biomass is associated with increasing elevation and decreasing salt concentration. All four levels of three classification systems, the Soil Taxonomy (ST), the Hungarian classification (HU) and the WRB were evaluated in a 100 m regular grid of 85 profiles for their applicability for biomass estimation (using 10-year average NDVI as proxy) and their correlation with ground elevation. NDVI values reflecting soil formation chronology (from the least to the most developed soils) were found on the first (least detailed) level of the classification systems. By analyzing the aspects of practical applicability, mainly at the detailed levels 3 and 4, HU performed the best in terms of class separability, WRB showed the most homogeneous classes, HU provided the closest correlation with elevation; while ST operated with the lowest number of classes, and, consequently, had a lower level of homogeneity and weaker correlation with elevation. Both HU and WRB performed well in most aspects, but the latter showed greater homogeneity. WRB had twice as many classes as HU and four times as many compared to ST; thus, their homogeneity increased accordingly.
The implementation of a soil classification without profound tests might result in counterproductive classes in terms of class separability, homogeneity of classes, correlation with environmental parameters, and parsimony of classes.
The results are published in Geoderma: