SPATIAL VARIABILITY OF SOIL PROPERTIES IN SELECTED ILLINOIS MOLLISOLS

We investigated two areas of contrasting spatial variability in east-central Illinois using grid sampling to determine the composition of soil map units and the variability of significant soil properties within map units. We also evaluated the effectiveness of the modern soil survey legend and maps in addressing the soil variability in the two different landscapes. The surface characteristics of the sample sites showed extremely high variability, especially in slope and aspect with CVs of 93.5% and 83.6%, respectively. The depth of the A horizon showed high variability with CV of 35% for both areas. The depth to reducing conditions was highly variable with a CV of 34.6%, while the depth to discontinuity was 48.4%. Characteristically the degree of color variability increased in the order, hue, value, and chroma, regardless of where in the profile it occurred. The A horizon showed very high variability in sand content with a CV of 74.4%, but silt content and clay content both showed moderate variability with CVs of 22.9% and 23.6%, respectively. Organic C content showed moderate variability with CV of 15% and was more variable in the highly undulating landscape, probably due to redistribution by erosion. Soil reaction in the B horizon showed a low degree of variation with CV of 9.5%. Although the variability of each soil property within the eight map units measured by coefficient of nondetermination showed less variance than the total variance present in both sample areas, most of the variation of soil properties was still left within map units. Tolerance limits of 75% set for the various map units indicate that some of the map units contained less variation of soil properties than others. This approach should help in refining the cartographical separations by identifying potential units that need reduction in variation of selected properties. Based on evaluation of soil property variability, we found that although the existing soil survey legend and maps are good, they could be improved substantially by considering the within-map unit variability of significant soil properties. We recommend that in future soil surveys, an initial soil property variability study be conducted using grid sampling to aid in a comprehensive development of the soil survey legend which will help reduce the within-map unit variation. © 1990 Williams & Wilkins.