Scale-Dependent Limitations of the Mehlich 3 Method in Paddy Soils in Japan: Toward Standardized Soil Testing

Abstract

The Mehlich 3 (M3) method is one of the most widely used universal extractants for multielement soil testing of nutrient availability, yet its performance across spatial scales that encompass diverse pedogenic contexts remains unclear. To assess the comparability of the Mehlich 3 method with conventional extraction methods across contrasting pedogenic contexts, M3 extraction was compared with conventionally extracted nutrient contents for 10 elements (Ca, Mg, K, P, Mn, Zn, Ni, Fe, Cu, and Si) in paddy soils collected across three spatial scales (local, regional, and national) in Japan. Regardless of spatial scale, the M3 method showed comparable extracted amounts to conventional methods for Ca, Mg, K, Mn, Zn, and Ni, thereby confirming its reliability as a universal extractant. Regarding P, the extracted amounts of M3 were about half those of Bray 2 and about twice those of Truog. Furthermore, the correlation coefficient between M3 and these conventional methods for P decreased at broader spatial scales, including soils with variable amorphous Fe/Al (hydr)oxide content. The present results demonstrate that the M3 method provides robust and scalable estimates for most nutrients, while also revealing that its performance for P is sensitive to pedogenic variability. This highlights the importance of incorporating pedogenic context when defining M3 applicability domains, particularly for P assessment, thereby improving the comparability of soil testing results across small- and large-scale soil monitoring frameworks.