Laboratory and field measurement of magnetic susceptibility of Japanese agricultural soils for rapid soil assessment

Measuring volume magnetic susceptibility (κ) of soil allows rapid soil assessment. In this study, we measured the κ value of agricultural surface soils in Japan at national and farm scales in order to evaluate the scale-dependent relationship between soil κ and other soil properties and to examine the repeatability of the κ value measured in the laboratory and in the field. A handheld field sensor was used to measure two sample sets: (1) 164 samples collected throughout Japan (national scale), and (2) 246 samples collected from paddy fields in Takatsuki and 117 (39 sites × 3 times) samples collected from paddy fields in Mifune (farm scale). Laboratory measurements showed that the coefficients of variation of soil κ were 112% for the national scale samples, 13% for the Takatsuki samples, and 33–37% for the Mifune samples. The large variation at the national scale was due to several positive outliers, which were classified as Vitric or Silandic Andosols by the WRB classification. Repeated measurement of samples enabled estimation of the within-sample variation in κ, and the data was considered unreliable if the coefficient of variation exceeded 10%. The correlation analysis using the reliable data indicated that at the national scale, the κ value was most positively correlated with structural Fe (total Fe minus Fed) followed by Alₒ and the andic properties (Alₒ + 1/2Feₒ). In contrast, for the Mifune samples collected at the farm scale, the κ value was most positively correlated with sand content followed by K saturation percentage. Field measurements of soil κ at the 39 Mifune sites indicated that the coefficients of variation of κ at the same site were almost all below 10%. The within-site variations were due probably to the incomplete contact between the soil surface and the sensor. Our results suggest that measuring κ of agricultural surface soils in Japan can help to distinguish Vitric or Silandic Andosols from other soils, and to estimate the κ-related soil properties at the farm and national scales.