Key points are not available for this paper at this time.
Soil maps derived from random or grid-based sampling schemes are often an important part of precision cropmanagement. Sampling and soil analysis to derive such maps require a large investment of both time and money. Aerialphotos have been used as a soil mapping aid for years. Studies have shown such an approach can be useful for definingmanagement units in precision farming, but these studies are often limited to a single field, not an entire farmingoperation. In this study, multispectral airborne green, red, near infrared (NIR), and thermal and satellite (SPOT andLandsat TM) data were used to derive soil textural class maps for 350 ha of a 770 ha research and demonstration farm inMaricopa, Arizona. These maps were compared to soil textural analysis results from samples in the top 30 cm of the soilprofile at an approximate grid spacing of 120 m. Differences in tillage, residue, soil moisture, etc. between fields limitedthe accuracy of spectral classification procedures when applied across the entire study area. However, using spectralclassification procedures on a field-by-field basis, it was possible to map areas of soil textural class with reasonableaccuracy. These results are specific to the study area and may not apply at other locations due to the numerous factorsthat can contribute to a soils spectral response. Classification procedures were also used with vegetation present over thestudy area later in the season. Resulting vegetation classes may be helpful in deciding if soil classes impact cropdevelopment enough to warrant different management practices.
Barnes et al. (Sat,) studied this question.