1997 Details

• The project site was divided into four quadrants (two corn and two soybeans) of approximately eight to nine acres each. Corn was planted at 32,000 and 36,000 seeds per acre, and soybeans were drilled (200,000) and planted in 30-inch rows (182,000).
• Soil sampling was done on both 0.62-acre grids and 1.1-acre grids (and by soil type in quadrants) by major soil mapping units per 10-acre quadrant and by 20-acre composite samples. Sampling involved checking potassium, phosphorus, and nitrate levels; soil pH; and nematodes.
• The late spring nitrate test and the fall stalk nitrate test were taken by 1.1-acre grids in the corn acres to monitor nitrogen (N) use.
• Weekly scouting for crop growth, weeds, insects, and diseases was conducted throughout the growing season.
• Yields were calculated by using a combine monitor, sample load-scale weights on the combine, and elevator scale weight per quadrant.
• A comparative statistical analysis of the various layers of information was completed to determine the correlation to yield data and limiting factors and to make recommendations for the future.
• Using the Iowa Crop Database software, returns to management and profit were calculated.

What we learned the first year

What did we learn this first year and what new questions were raised?

0. The geographical information systems (GIS) equipment is very accurate, allowing us to get back to the same point in a field.
1. Expensive electronic equipment will not function properly with loose connections, discharged batteries, or dust or pollen on the computer screen.
2. Field variation is not in squares or particular-sized grids. Statistical extrapolation between point samples did not increase the accuracy of predicting variation.
3. Differences in yield did not correlate consistently with pH and soil fertility variation.
4. The obvious limiting factor in soybeans was the level of white mold infection. The limiting factor for corn production was not obvious.
5. Calibration of harvest monitors is imperative and particularly difficult in small or odd-shaped fields.
6. Scouting data were available in a timely manner. Weeds, insects, and N problems could be corrected before they caused economic yield reduction.
7. Yield correlation data are just now available--too late for fall fertilization decisions.
8. Human interpretation, decision making, and management are still necessary. Without such input, you just have a lot of fancy maps.
9. Variable-rate application by grids may increase variability in the field. With fertilizer or lime, these changes would be irreversible in the short term. If the point sample was not representative of the grid, problems may be compounded rather than corrected.

Conclusions

Should farmers invest in harvest monitors and global positioning systems (GPS) /GIS equipment? Should farmers consider hiring companies to do grid sampling and variable-rate applications? Certainly a harvest monitor would document yield variation due to poor drainage, weed, or disease pressures, and hybrid and variety performance. Over time, these tools will provide a valuable record of crop performance.

In most cases the answers will be based on individual priorities. Where will the return on investment be greatest? In GPS/GIS equipment; a better planter; an applicator control; a weigh wagon scale for grain and feed; hiring a crop consultant; or a computer record-keeping program?

Precision agriculture tools can be just the latest toy--or a valuable tool to improve crop management.