Evaluation of the accuracy of a central Iowa soil survey and implications for precision soil management

The movement towards precision agriculture has led to calls for soil maps that are more detailed and accurate than those offered in standard NCSS soil surveys. Studies have shown that soil variability can be greater than depicted in soil surveys; in fact, delineations that contain at least 50% of the soil mapped are considered satisfactory for soil survey purposes. Lacustrine plains are relatively fiat and often have parent materials with uniform properties. Because soils are usually mapped using soil-landform relationships one might expect soil maps in these areas to be less accurate than average; it is difficult to delineate between map units using soil-landform relationships in such subtle landscapes. We grid-mapped a field containing lacustrine-derived soils in central Iowa and used the grid to evaluate the soil survey for accuracy. Two major and two minor soils, as determined by the area they occupy in the field, were present. For the field as a whole, the two major soils were correctly identified by the soil survey at least 63% of the time. The two minor soils were correctly identified 33% of the time or less by the soil survey. Large-scale soil mapping is expensive because of the time involved to create them in the field and in the office. Therefore, it is only economically beneficial to produce a detailed map if the map leads to significant alterations in the way a field is managed. In fields that may have uniform soil properties, it may be more cost-effective to conduct a 'reconnaissance' survey first and then decide if more detailed mapping is required.