Marysville rice grower Charley Mathews, Jr. was one of the first California rice producers to try the new yield mapping technology when he purchased a yield monitor and mapping software in 1997. The results of mapping yield in a few fields that year demonstrated two things.

The first was that a rice field that looks uniform from the windshield can have a surprising amount of yield variability. The second was that mapping this variability raised more questions than it answered. In many of the fields some areas were clearly not yielding up to their potential.

In an effort to find out why, Mathews and other Sacramento Valley rice growers have teamed up with University of California researchers to examine the factors underlying spatial variability in yield, using research support from the Rice Research Board and the John Deere Corp. Not surprisingly, they have found that there are almost as many possibilities as there are rice fields.

One of the best ways to analyze the data that comes from a yield map is through a geographic information system, or GIS, which can be used not only for visual comparison of maps, but also for statistical analysis. During 1998, Mathews collected yield data from a problem field. The field generally yielded well, but had three major bad areas, in the southwest corner, the southeast corner, and near the middle of the west side.

The pattern of low yield bears a strong resemblance to the pattern in a false color infrared aerial photograph of the field taken in April. A false color image shows reflected infrared radiation as red. Since healthy vegetation strongly reflects infrared radiation, red areas in the image are those with a good stand. It is evident that the low yielding areas had not established a good early stand.

Examining the data for causes of the poor early stand led to the conclusion that the primary cause was variability in water depth. It is important to note that it was not the depth of the water that caused the problem but rather the fact that the water depth was highly variable, so that different parts of the field were at different levels of development. This effect was probably exacerbated by the very cool early spring in 1998, so that rice in deeper water emerged later than rice in shallower water and stayed behind throughout the season.

Up to this point, the analysis was primarily a visual comparison of maps, but now the analytical power of the GIS was added to the effort. Alvaro Roel, a graduate student at UC Davis, carried out an economic analysis of the data. He first converted the grain yield map to an economic yield map.

Although yield varied across the field, inputs were applied uniformly, so the costs were the same at each point. The break-even yield value was subtracted from the grain yield and the resultant number was multiplied by the price per pound of rice. This resulted in the profit map. This map has superimposed on it the boundaries between the high and low elevation areas of the field.

By computing the difference between mean economic yield in the high and low areas and multiplying this by the total low area, Roel could estimate the return on a laser leveling operation.

The laser leveling was in fact carried out in the spring of 1999. Figure 5 shows the yield maps of the two years side by side. It is evident that much of the variability in yield that was present in 1998 was gone in 1999. Not all of this reduction can be attributed to the laser leveling, of course, since other factors such as differences in weather may have also been important.

There is still much to be learned about the use of site-specific management tools in rice production. This example--and others like it--shows that by keeping the focus on the bottom line, these tools can be used to improve economic productivity.