Evaluating the detection limit of organic matter using point and imaging spectroscopy

The commonly used optical technique for estimating organic matter in soils is based on a chemometric approach. Many studies have examined the correlation between spectral response and organic matter content, but none have investigated uncertainties based on unequal spectral responses in various soil types or spectral detection limit. The aim of this study was to systematically examine the spectral responses of five different soils with increasing amounts of organic matter to evaluate detection limits in point (ASD) and image (HySpex) domains. In addition, we evaluated the contribution of clay content and soils' initial organic matter on the assessed detection limit for each sensor. Large spectral variations were found between soils with the same organic matter content and between the two sensors when calculating the detection limit. Thus, applying a generic prediction model of organic matter using all soil types results in a rough estimation due to the spectrum affiliation with a certain class and not due to correct organic matter analysis. The importance of this study lies in showing that the use of spectroscopy for spectral-based organic matter detection should be practiced with caution, by highlighting the variant and inconsistent effects of organic matter in soils, and discussing the problematic assessment of organic matter when using a chemometric approach.