Comparative verification of Mehlich 3 soil analysis methods in Northern Mozambique using microwave plasma–atomic emission spectrometry

Abstract

ABSTRACT Regular testing of nutrients in soil samples is important for understanding soil fertility. This study investigated a combination of the Mehlich 3 extraction method and Microwave Plasma-Atomic Emission Spectrometry (MP – AES) for testing the nutrient status in soil to establish a labor-saving and cost-effective analytical method in Northern Mozambique. Available P, exchangeable cations, and micronutrients in the Mehlich 3 extract were measured using conventional methods and compared with those quantified using MP – AES to determine its accuracy. Linear regression analysis implied a high correlation between the Available P extracted via Mehlich 3 and quantified using MP – AES and that extracted via Mehlich 3, Bray 1, Olsen, and Mehlich 1 and quantified using UV – Vis. Moreover, a high determination coefficient (R2 = 0.920) was observed between Mehlich 3 quantified using MP – AES and that quantified using UV – Vis spectroscopy with a low root mean square error (RMSE = 9.245), MP – AES may have accuracy comparable to conventional methods. Linear regression of the exchangeable cations extracted using the Mehlich 3 method and quantified using MP – AES and those extracted using the 1 M NH4OAc method and quantified using ICP – AES revealed that the coefficients of determination of Ca, Mg, and K were 0.952, 0.878, and 0.826, respectively. The coefficient of determination for Na was small at 0.101. For the micronutrients, Mehlich 3 extraction with MP – AES quantification method showed a correlation with the conventional DTPA method; specifically, a strong correlation was observed for Cu and Zn. Although the correlation for Fe and Mn was also high, it was less accurate than that for the other elements. The method used in this study is considered effective for estimating field variability. To convert to conventional methods, variation in available elemental content between sites with different local soil properties should be considered.