Applicability of Sequential Chemical Extraction in Determining Phosphorus Transformation during Soil Development Under Semi-Arid Climate.

The Walker and Syers (WS) model, developed based on chemical sequential extraction (CSE) analysis, is widely accepted for describing phosphorus (P) transformation in soils with increasing substrate age. However, CSE provides operationally-defined P fractions and thus the accuracy of the CSE-based WS model is questionable. Here, the Hedley CSE results were compared with P K-edge XANES analysis regarding P transformation in surface soils of the San Francisco Volcanic Field soil chronosequence (1, 55, 750 and 3000 kyr) located in northern Arizona. Results show that P transformation obtained from the CSE approach was consistent with the WS model; however, the P XANES linear combination fitting (LCF) analysis indicates that the Ca‑P fractions in the two younger soils were highly overestimated while Fe/Al-P was underestimated by CSE. To further explore the discrepancies of the two methods, LCF analysis of the difference XANES spectra between the residuals of two successive extraction steps show that the P removed by 1 M HCl extraction was almost 100% Ca-P in the two older soils, consistent with the CSE assignment, while it included up to 66% Fe-P in the two younger soils. Mössbauer and Fe EXAFS spectroscopy further revealed that the two younger soils contained abundant non-crystalline Fe oxides while the older ones did not. The dissolution of the non-crystalline Fe oxides by 1 M HCl, hence releasing the P bound to these minerals, accounted for the overestimation of the Ca-P fractions in the two younger soils. These results indicate that the accuracy of CSE in determining P transformation depends on soil mineralogy that is affected by soil developmental stages. This work suggests that the widely‑accepted WS model needs re-evaluation using direct soil P speciation approaches, such as P XANES spectroscopy.