Analytical data for 10 major oxides (Al2O3, CaO, Fe2O3, K2O, MgO, MnO, Na2O, P2O5, SiO2 and TiO2), 16 total trace elements (As, Ba, Ce, Co, Cr, Ga, Nb, Ni, Pb, Rb, Sr, Th, V, Y, Zn and Zr), 14 aqua regia extracted elements (Ag, As, Bi, Cd, Ce, Co, Cs, Cu, Fe, La, Li, Mn, Mo and Pb), Loss On Ignition (LOI) and pH from 3526 soil samples from two continents (Australia and Europe) are presented and compared to (1) the composition of the upper continental crust, (2) published world soil average values, and (3) data from other continental-scale soil surveys. It can be demonstrated that average upper continental crust values do not
provide reliable estimates for natural concentrations of elements in soils. For many elements there exist substantial differences between published world soil averages and the median concentrations observed on two continents. Direct comparison with other continental datasets is hampered by the fact that often mean, instead of the statistically more robust median, is reported. Using a database of the worldwide distribution of lithological units, it can be demonstrated that lithology is a poor predictor of soil chemistry. Climate-related processes such as glaciation and weathering are strong modifiers of the geochemical signature inherited from bedrock during pedogenesis. To overcome existing shortcomings of predicted global or
world soil geochemical reference values, we propose Preliminary Empirical Global Soil reference values based on analytical results of a representative number of soil samples from two continents (PEGS2).
Keywords: Regolith; Critical Zone; Geochemistry; Major elements;Trace elements; Global soil composition