Use of precision agriculture techniques in Kentucky during the past several years has generated interest in how to soil sample a field for use in programming computer-driven, on-the-go, variable rate fertilizer spreaders (VRS). The advantage achieved by VRS is related directly to variability of soil test (ST) values within a specific field and the accuracy of how they represent the field. Since variability of ST values commonly exists on a small scale, a very intensive sampling procedure (grids of one acre or less in size) would be required to accurately describe the nature and extent of such variability within a field. The cost of sampling and analysis on such a scale would be prohibitive to most commercial producers. For this reason, many fertilizer dealers offering VRS to customers recommend sampling on a 2.5 acre grid (330x330 ft), and use this 2.5 acre unit as the basis for VRS within a field. Use of this procedure to vary fertilizer rates within a field is based on the assumption that variability of ST values within each 2.5 acre block is less than that for the field as a whole. And further, this assumes that ST values are fairly uniform across the swath width of the spreader (about 60 ft) and along the 330 ft pathway of the spreader as it is driven through each 2.5 acre block. Both these assumptions are questionable. Wells (1) reported variation in ST phosphorus (STP) of nearly two-fold across and along 40 ft wide spreader swaths in a 3.4 acre field which was intensively sampled in Shelby Co., KY. Recommendations for phosphate fertilizer rates among the 162, 8 x 20 ft blocks sampled in that study varied from 0 to 110 lbs/A. If the entire 3.4 acre area had been fertilized with a uniform rate based on results from one composite sample taken randomly from within the 3.4 acres, the entire area would have received 80 lbs P2O5/A. Application of the uniform 80 lb rate to the 3.4 acres as compared to the rate which would have been required for each of the 162 areas sampled within the 3.4 acres, would have resulted in only 31% of the area receiving the correct rate. Additionally, 39% would have been underfertilized and 30% would have been overfertilized. Such variability within a 3.4 acre block would not likely be overcome by use of a VRS programmed with the capability to vary rates every 2.5 acres.
Wells, Kenneth L. and Dollarhide, James E., "Precision Agriculture: A Field Study of Soil Test Variability And Its Effect on Accuracy of Fertilizer Recommendations" (1998). Soil Science News and Views. 8.