a All authors are at U.C Davis, except for Cahn and Kearney, who are with U.C. Cooperative Extension in Sutter-Yuba and Yolo counties. Author for correspondence: Stuart Pettygrove Department of Land, Air and Water Resources University of California, Davis, CA 95616 Email: gspettygrove@ucdavis.edu.

Introduction

In the fall of 1995, a team of U.C. scientists and Cooperative Extension specialists and advisors began working with Yolo County grower Tony Turkovich to attempt to relate within-field variation of crop yield and quality to variations in soil, pest pressures, plant tissue nutrient content, etc. The team's intent is to relate environmental variables (such as soil drainage class) and manageable factors (irrigation, fertilizer) to crop yield and quality using relatively low-cost information obtained through aerial photography and yield mapping. A key question is: Can variability and its causes be mapped without over-reliance on more expensive information such as grid soil and plant sampling?

The project is supported by a two-year grant from the California Department of Food and Agriculture Fertilizer Research and Education Program. Also, U.C. is contributing a significant amount of technician time. Team members bring a wide range of expertise to the project including in crop modeling and plant physiology, agronomy, irrigation, soils and plant nutrition, engineering, remote sensing, and geographic information systems.

Procedures

Results and Conclusions

The net grain yield in this 77-acre field was 2,944 lb./acre - less than half of a typical "good" wheat yield in the southern Sacramento Valley. The average yield recorded by the yield monitor was 2566 lb./acre. The 13% discrepancy between yields obtained from the truck weights and the yield monitor was probably due to insufficient calibration of the latter. The great variability of the yield across the field is seen in Figure 1. Yield in the southwest corner exceeded 4,000 lb./acre but was less than 1,000 Lb./acre in the north central area of the field. Additional smoothing and processing of yield data -- for example, to remove data noise caused by short-range fluctuations in the flow of grain through the combine-- is still required before an accurate characterization of yield variability can be obtained.

The main factor contributing to low yield was saturated soil conditions resulting from poor soil drainage and heavy winter rains. Apparently, growing the wheat on five-foot beds did not compensate for the slow drainage characteristics of the soil. Yield was highest in the southern one-third of the field where the soil was lower in silt content and higher in sand (Fig. 2). The lowest grain yields were observed in the northeast quarter of the field where the surface soil was higher in silt and clay and where a restricting layer (> 50% clay) was present at a depth of three to five feet. This area of the field also displayed the darkest color in the December bare soil aerial photo (not shown), taken shortly after a rain.

A second cause of low yields was competition from grassy weeds. The density of the weeds varied greatly across the field. In some areas, there were no weeds. In a few areas, there was almost no wheat, only weeds. Weed ratings shown in Fig. 3 were obtained by one person walking the entire field. The weed species that caused the biggest problem were wild oats, canarygrass, and ryegrass. Because these species are equal in height or taller than the wheat and possess a different color seed head, they are easily seen in the May aerial photograph (not shown). We plan to further analyze these data. It may be possible to produce a weed density map from the aerial photograph and -- by comparing that to the yield map -- determine the grain yield and economic loss caused by lack of weed control.

An unanticipated finding was the presence of dark green (darker red in the color IR) streaks in the May aerial photograph. These streaks run in a north-south direction and are about 70 feet apart. We believe the streaks are the result of non-uniform aerial application of N fertilizer topdressing on February 25. The streaking is also faintly visible on the March 8 aerial photograph and on the yield map. Even in the greatly degraded version of the yield map produced for this summary paper (Fig. 1), there is a north-south pattern visible in a few places, although without further analysis we cannot be certain that it is the same pattern seen in the aerial photographs. We will analyze the yield monitor data for the presence of an east-west cyclic pattern in grain yield and moisture content and will compare that to the color pattern in the aerial photograph. This example shows the tremendous potential for using within-field variability - in this case caused by an unintended non-uniformity in fertilizer application - to conduct experiments that could otherwise only be done with expensive small-plot experimentation. A further benefit of this whole-field approach to research is that the effects of fertilizer N can be observed across the range of conditions present in the field. A single, compact small-plot experiment would probably cover only one soil type; in fact, the researchers would try to place the small plots in the most uniform, problem-free area of the field. This example also shows how yield mapping technology used alone or in combination with aerial photography can enable a grower to conduct his or her own experiments at a relatively low cost.

The authors acknowledge assistance from Kurt and Dean Wesley, Key Agricultural Macomb, IL, U.C. Davis staff members Victor Huey, Deng Jiayou, and Kent Kaita, the U.C. DANR Analytical Laboratory, and Button & Turkovich Farms.