Site-Specific Application of Fertilizer in Soybeans

Linda Benedict, Moore, Steven H., Wolcott, Maurice C., Bechtel, Amos  |  8/4/2009 1:56:01 AM

Steven H. Moore, Amos I. Bechtel, Robert G. Downer, Maurice C. Wolcott and Michael L. Tarpley

New technology using computers and satellites has made it possible to measure the variability of nutrients within a field and vary the rate of applied fertilizer based on need. A two-year study of site-specific application of fertilizer in a soybean field at the LSU Agricultural Center’s Dean Lee Research Station in Alexandria involved three tools: a global positioning system (GPS), geographic information systems (GIS) and variable rate technology (VRT).

The global positioning system is operated by the U.S. Department of Defense and involves 24 satellites that transmit radio signals to earth. Receivers on the ground collect these signals to determine longitudinal and latitudinal coordinates. Most receiving systems used in agriculture can calculate location within a few feet. Vehicles, tractors and combines outfitted with receivers are able to determine location at any point in a field.

Geographic information systems are computer software programs used to map and analyze geographic data, such as nutrient concentration in a production field. GIS software matches data with exact coordinates to create a map. The maps provide useful information to producers, such as how organic matter, phosphorus concentration or yield may vary within a production field. A prescription map may be constructed from a soil nutrient map to show the amount of input needed at any site.

Variable rate technology combines GPS and GIS with application hardware to vary the rate of input according to the prescription map. For example, a fertilizer spreader with VRT capability includes a receiver to collect signals from satellites, a computer equipped with GIS and a prescription map of the amount of fertilizer needed at specific sites in the field, and a series of controllers that operate according to computer instructions.

Mapping soil fertility

The study was conducted in 1997 and 1998 on a predominantly Norwood silt loam soil. The test area was about 162 acres. Plot sizes were either 8 or 11 acres. The three treatments in the tests were: fertilizer applied at a uniform rate, fertilizer applied at a variable rate based on a 1-acre grid sampling pattern and fertilizer applied at a variable rate based on a 2.5-acre grid sampling pattern. Soybeans were planted both years in 38-inch rows.

Pettiet Ag Services of Leland, Miss., did soil nutrient maps in the spring of 1997. At the time, no company in Louisiana had begun mapping soils for nutrient content. Samples were collected using both a 1.0-acre and a 2.5-acre grid. A composite sample for the field was collected also. The cost for mapping the test using a 2.5-acre grid pattern was $8 per acre, and the cost for the 1.0-acre grid pattern was $20 per acre.

The boundaries of the field were first determined by driving the perimeter of a field in an ATV equipped with a receiver and GIS. A grid pattern was then overlaid on the field, and a sampling point within each grid was randomly assigned. Samples were collected by driving to each point on the map and pulling soil cores.

Samples were analyzed for phosphorus, potassium, magnesium, calcium, sulphur, zinc, boron, cation exchange capacity, organic matter and pH. Nitrogen is almost never tested for in fields planted in soybeans. Soil test values were entered into the computer, and the data were extrapolated to construct color-coded soil nutrient maps for each grid sampling scheme. Since phosphorus, potassium and sulphur were found deficient, the prescription maps were constructed to show the amount of fertilizer needed at every location in the field to bring these nutrients to desired levels.

Variable rate application

Phosphorus, potassium and sulphur fertilizers were applied at variable and uniform rates using a Terra-Gator (Model 1903) manufactured by Ag Chem Equipment Company, Inc. The Terra-Gator was equipped with a GPS receiver and a Falcon control system (GIS and VRT controllers). The Terra-Gator had four fertilizer bins and applied the three fertilizers in one pass. The spread width was about 20 rows, and the truck ran at about 19 miles per hour, allowing an application time of about an acre a minute. Fertilizer was applied to overwintering beds on April 15, followed by lifting the middles to cover the fertilizer material.

Crop culture

A John Deere vacuum planter was used to plant Asgrow A5885 in 1997 and Pioneer P9611 in 1998. The soybeans were harvested with a John Deere 9400 field combine equipped with a Green Star Yield Monitoring System. Yields were recorded using scales at the elevator where the grain was sold.

Soil nutrient maps for phosphorus and potassium showed the nutrient concentration generally increased from west to east. Finer differences in nutrient concentration were observed in the maps using a 1.0-acre grid sampling pattern, compared to the 2.5-acre grid sampling pattern. More “islands” appeared in the maps based on 1.0-acre grid sampling patterns, but both grids showed similar patterns indicating that the 2.5-acre grid sampling pattern was able to depict soil nutrient variability. Overall, the prescription maps called for less fertilizer in the variable-rate applications than in the conventional uniform application treatment.

Little yield difference

Yield data from the two-year study indicated no significant differences among the three fertilizer application methods. The yield response of soybeans to potassium and phosphorus fertilizers at this location is historically minimal or nonexistent. Assuming no difference in yield, the economic analysis comes down to application costs. Fertilizer costs used in the economic analysis here are $160 per ton for 0-0-60, $230 per ton 0-46-0 and $330 per ton for Sulf90. The soil mapping and testing cost was $20 per sample, and one sample was required for the entire field when uniform fertilization was practiced.

One sample and soil test was required for each 1.0-acre and 2.5-acre grid. Soil sampling is generally done once every three years, so the cost of sampling and soil testing is allocated over a three-year period using a 9 percent cost of capital. Since no significant differences in crop yields were found among the three treatments, total revenue per acre is the same for all. Differing profitability among treatments is reflected solely by the difference in cost associated with the sampling grid size, quantity of fertilizer applied and fertilizer application. These costs are shown in Table 1.

For the two years of this study, uniform or whole field sampling and fertilizing was the lowest cost system, with annual fertilizer, sampling, testing and application costs of $30.20 per acre. The costs for the 2.5-acre and 1-acre grid systems were $32.03 and $34.51 per acre, respectively. In this case, uniform fertilizer application increased profitability by $4.31 per acre over the 1-acre grid and by $1.83 per acre over the 2.5-acre grid. Note that in this study only fertilizer cost is being examined. Additional benefits resulting from the information gained through more intensive soil sampling have not been considered. These benefits might include reduced soil-applied herbicide costs made possible by varying application rates to match soil properties or by variable seeding rates that could potentially increase crop yields or reduce seed costs. Note, also, that greater benefits to grid sampling may be achieved when used on crops requiring larger fertilizer applications such as corn or cotton. More research is needed.


Appreciation is acknowledged to the following agencies that contributed to making this research possible: Louisiana Soybean and Grain Research and Promotion Board, International Teaching and Research Foundation, Ag Chem Equipment Company, Inc., and Potash and Phosphate Institute.

Steven H. Moore, Professor, Dean Lee Research Station, Alexandria, La.; Amos I. Bechtel, Assistant Professor, Agricultural Economics Department, LSU Agricultural Center, Baton Rouge, La.; Robert G. Downer, Assistant Professor, Experimental Statistics Department, LSU Agricultural Center, Baton Rouge, La.; Maurice C. Wolcott, Research Associate, and Michael L. Tarpley, former Research Associate, Dean Lee Research Station

(This article was published in the summer 1999 issue of Louisiana Agriculture)

Rate This Article:

Have a question or comment about the information on this page?

Innovate . Educate . Improve Lives

The LSU AgCenter and the LSU College of Agriculture