In 1997 a Precision Agriculture Demonstration Project was established near the Iowa State University (ISU) Northeast Iowa Research Farm at Nashua. The project was developed to provide producers and service providers with practical recommendations to realize the potential benefits of this new technology. The unique emphasis is to make cropping decisions based on integrated crop management principles and the information gathered by using global positioning system-geographic information system (GPS-GIS) technologies.

The 40-acre project site is divided into quadrants (two for corn and two for soybeans). The only planned variables are planting rate in the two corn quadrants (32,000 and 36,000) and row spacing per seeding rate in the two soybean quadrants (10-inch drill/200,000 and 30-inch row/182,000). Observations from year 1 were reported in the Precision Agriculture Special Edition (Spring 1998) issue of the ICM newsletter.

In year 2 we checked pH, phosphorus (P), and potassium (K) levels within each 1.1-acre grid by using three different methodsóa point sample from the center of the grid, a sample from a random point within the grid, and a composite sample of the grid. Rainfall was monitored at nine points in the project site.

Planter accuracy was monitored using a computerized measuring wheel called a Space Cadet that measures population density and plant spacing. The late spring nitrate test (LSNT) and the fall stalk nitrate test (FSNT) were used to monitor nitrogen (N) levels. Field scouting was conducted to document any weed, insect, or disease problems. Yields were calculated by quadrant by using the combine monitor, strip tests by using the load cell scale on the research farm combine, and scale weights from the elevator. Statisticians at Iowa State University and in Kansas compared the various layers of information with yield to try to establish correlations and identify limiting factors.

Observations and conclusions

1. Our experience with differences in soil test results from various sampling methods of an area reflects the findings of research conducted at Iowa State University and elsewhere.
2. It is difficult to get a "representative sample" of the whole. Whether we're sampling soil (a 1-lb soil sample out of a 2 million lb/acre furrow-slice), plant populations, lodging, or percent barren (from 1/1,000th of an acre), grain moisture (a handful from a grain tank), or some other variable, there is considerable room for variation and misrepresentation. There was considerable variation between and within grids that could be partially explained by sampling error.
3. Rainfall differences within the 40 acres exceeded 2.4 inches during the growing season.
4. Scouting data were available in a timely manner. Weed, insect, and N deficiency problems could be corrected before they caused economic yield reduction.
5. Anhydrous ammonia toolbars and dry fertilizer/lime and manure spreaders currently in use may compound soil test variability because of nonuniform spread patterns. Calibration, maintenance, and adjustment of planting, spreading, and spraying equipment are critical to implementing precision agriculture systems.
6. Combine monitor data at specific points within a field, from small plots, or from odd-shaped fields, may not be sufficiently accurate or reliable and should be considered suspect.
7. Dust, pollen, moisture, and adverse temperatures may affect the accuracy and reliability of GPS-GIS equipment.
8. By using elevator scale weights, we determined that the project corn averaged 193.96 bu/acre and the soybeans 65.9 bu/acre. Statistical analysis comparing yield to the various layers of information collected did not identify any strong correlations or obvious limiting factors. The question is raised, "How precise do you need to be?" Our 1998 experience would suggest that in a year with favorable growing conditions, most everything works.
9. Human interpretation and decision making are still required. Precision agriculture technology does not replace management. Because of the voluminous amounts of data that can be collected, making appropriate management decisions may actually be more difficult.
10. We need to frequently remind ourselves that the primary purpose of our project is to demonstrate integrated crop management (ICM) decision making. We are using GPS-GIS technologies to expedite information gathering for that decision-making process. ICM is the "end" and precision agriculture technologies are a "means to the end." ICM is a stand-alone conceptóthe GPS-GIS technologies are not. We are confident that most farmers can benefit from ICM decision making. Based on questionnaire responses from Precision Agriculture Field Day participants, investments in precision agriculture technologies are difficult to justify for many farmers.