Distribution law and evaluation of chemical elements contents in soils below the detection limit

There are samples in soils, where the content of any element falls below the limit of detection (for example, X-ray fluorescence analysis). To reveal a very low element content, we can achieve by regression analysis to obtain the dependence of its content from a common chemical analogue in profile. The most effective approach is the application of power regression equation, connecting low-Clarke lanthanides contents with high-Clarke lanthanides contents and low-Clarke bromine content with the high-Сlarke halogen chlorine in soils.

There is a gradual transition from the normal distribution of high-Clarke lanthanides to lognormal distribution in low-Clarke lanthanides. The power regression approach allowed us to separate the regression relative error from the total metrological error. Estimation of the content of dispersed elements in the soil below the detection limit is limited by the magnitude of the regression relative error, and an estimation of the regression relative error is important for the construction of geochemical relationships of the elements.