Modeling soil color and soil fertility relations

Soil color provides significant information on soil properties. Soil color frequently changes on cultivated semi-arid sloping landscapes, indicating differences in soil properties that affect soil fertility. This study aimed to determine the relationships between color variables (L*: soil brightness, a*: soil redness and b*: soil yellowness) and some basic soil properties on 155 soil samples representing soils on a semi-arid sloping landscape under a wheat cultivation. Soil properties were determined in the laboratory. Color properties measured with a spectrophotometer on both air-dry (L*-dry) and wet (approximately FC) (L*-wet) soil samples. The values of color variables and soil properties were graphed and relationships between them were modeled using most proper regression model. Among a*, b*, and L*, L* values were the most strongly correlated with soil properties and the correlations were the strongest when samples were wet. The CaCO3 content of the soils was the most significant variable associated with L*-wet (r=0.84). The relationship between L*-wet and CaCO3 content was described with a first-degree linear regression equation. The EC values were correlated relatively strongly with L*-wet (r =0.74), and the relationship was described with a third-degree polynomial equation. Clay, sand, and K content were correlated with L*-wet to a lesser extent, and silt, OM content, FC, WP and ASI were not significantly correlated. The extremely high CaCO3 content of soils would probably mask OM to show its association with L*. The results suggested that the water content of soil is an important variable in this relationship and that prediction of soil fertility from color variables may not be possible due to the high CaCO3 content.