Mapping Yield in Soybeans

Linda Benedict, Moore, Steven H., Wolcott, Maurice C.  |  8/4/2009 8:15:51 PM

Steven H. Moore, Maurice C. Wolcott and Michael L. Tarpley

Row crop producers have always had an idea of how yield varied in different parts of their fields. By looking at the crop before harvest and visually monitoring yield during harvest, they could see general yield patterns and speculate on why they occurred. But, because of new technology combining satellites and computers, the ability to picture yield patterns in a field and make diagnoses of the causes have been greatly advanced. A two-year study was undertaken at the LSU Agricultural Center’s Dean Lee Research Station to assess how this new technology can be used as a diagnostic and management tool.

Yield monitoring system

Yields were mapped in this study with a John Deere GreenStar Yield Monitoring System installed on a John Deere 9400 field combine equipped with a six-row header. The John Deere GreenStar Yield Monitoring System has five principal components: 

  • a GPS system that receives radio signals from satellites to determine longitudinal and latitudinal location
  • a mass flow sensor at the top of the clean grain elevator to measure the force exerted by grain exiting the elevator
  • a moisture sensor mounted on the return clean grain elevator
  • an auger speed sensor
  • a computer, located behind the operator’s seat in the cab of the combine, to synthesize data from the other four components. 

A yield measurement is made every 2 seconds when the header of the combine is down and engaged with the system turned on. The combine may operate for 250 hours before data need to be dumped from a PCMCIA card onto the hard drive of a desktop computer equipped with JD Map software. Instructions are in a manual that comes with the software to make yield maps.

Mapping yield

Soybean yields were mapped in a 162-acre soybean field in 1997 and 1998. The soil type is predominantly Norwood silt loam and is traditionally cropped to cotton. Asgrow A5885 was planted the first year and Pioneer P9611 the second. The soybeans were produced in 38-inch rows on raised beds. The first season was cool at the beginning but fairly normal afterward. The second season was extremely hot and dry for the region.

Raw yield data for 1997 are color coded in Figure 1. The average yield for the field in 1997 was 41 bushels per acre, according to elevator scales. The white gaps in the yield map are turn rows and a drainage ditch. The highest yields occurred mostly in the western portion of the field. Lowest yields were in the eastern part, and trouble spots are easily distinguished. The four rectangular cuts in the western two-thirds of the field are 24 acres each. The other two are 33 acres each.

Raw yield data may be contoured using extrapolation procedures to show broad yield patterns. This is a helpful procedure for distinguishing management zones. A trend in precision agriculture is to define management zones for individual treatment.

Yield patterns from the 1998 soybean harvest were similar to those in 1997 but showed additional features as well. In 1998, the average yield for the entire field was 40 bushels per acre, according to elevator scales. Temperature exceeded 105 degrees F during the season, which was relatively dry. The hot and dry conditions seemed to bring out new areas of the field where yield was markedly low.

Comparing yields

Twelve plots were harvested in 1997 where comparisons could be made between elevator totals and the yield monitoring system (Table 1). The calibration coefficient used to transcribe mass flow data into yield was periodically updated in the same field. In 1997, the difference in yield measurements between the GreenStar system and the elevator varied by up to 21 bushels per acre, although a difference of this magnitude was rare. All other differences were less than 10 bushels per acre. There was a difference of only three bushels per acre when measuring yield for the entire field.
 
Seventeen comparisons were made in 1998 (Table 2). One calibration coefficient was used for the entire field. The largest difference was 13 bushels per acre, with all others being 10 bushels or less. There was a six-bushel difference between the two measurements when considering the entire field.

The differences in yield measurements point to the need to use large, representative acreage when calibrating the yield monitor. It is a good practice to continually update the calibration, perhaps at the end of each day when elevator totals and acreage are known.

Using yield maps

Yield maps may be used as both diagnostic and management tools. Studying a yield map may help diagnose the cause of a particularly low or high yielding area by linking the information on the map with what is known about the field. Sometimes the cause of a low-yielding area is unknown. For example, the low-yielding area in the bottom center of to use large, representative acreage when calibrating the yield monitor. It is a good practice to continually update the calibration, perhaps at the end of each day when elevator totals and acreage are known. Using yield maps Yield maps may be used as both diagnostic and management tools. Studying a yield map may help diagnose the cause of a particularly low or high yielding area by linking the information on the map with what is known about the field.

Sometimes the cause of a low-yielding area is unknown. For example, the low-yielding area in the bottom center of to use large, representative acreage when calibrating the yield monitor. It is a good practice to continually update the calibration, perhaps at the end of each day when elevator totals and acreage are known. Using yield maps Yield maps may be used as both diagnostic and management tools. Studying a yield map may help diagnose the cause of a particularly low or high yielding area by linking the information on the map with what is known about the field. Sometimes the cause of a low-yielding area is unknown. For example, the low-yielding area in the bottom center of the field is believed to be caused by compaction; the cause of the low-yielding strip in the northeast section of the field is unknown. There are at least two approaches to determine the cause for low yield in this strip. One is to sample soil inside and outside of the low-yielding area. Comparison of these data may provide a reason for the yield difference. Another approach is to map other parameters affecting yield for the entire field, such as soil compaction or micro-topography related to drainage.

Additional parameters have been mapped for the field, including tissue nutrients in 1997, DRIS indices (nutrient balance) in 1997 and electrical conductivity in 1998. Correlation between yield and these parameters revealed how each was related to yield (Table 3). For example, 29 percent of the variability in yield could be accounted for by the variability of electrical conductivity (first column in Table 3). Furthermore, yield went up as electrical conductivity went down.

The correlation between phosphorus and yield was low. This does not mean phosphorus was unimportant, only that it was not a very high limiting factor. One-third or more of the variability in yield could be accounted for by the variability in potassium, suggesting that this nutrient often may have been a limiting factor.

The ability to correlate yield data with factors affecting production in individual fields provides a tremendous advancement in diagnostic ability for correcting yield-limiting factors. A second major economic advantage is that the solution need be applied only in the problem portion of the field. One caution in using correlation coefficients is that sometimes a parameter may not directly affect yield, but only be highly associated with a parameter that does.

Spatial stability of yield

Yield maps may provide an excellent tool to diagnose what happened last year. Their use for making management decisions, such as fertilizer recommendations, will be determined largely by the spatial stability of yield. Spatial stability measures the relative consistency in yield ranking of one location of a field with another from one season to the next. If there is no spatial stability, then yield maps provide no information for the future. If spatial stability is high, yield maps provide much information.

A limiting factor in soil grid sampling for site-specific fertilizer application is the cost. In addition, only one sample is usually collected for every 2.5 acres, which leads to the likelihood of missing variability in Louisiana soils. Yield maps often contain 200 or more measurements of variability at a relatively inexpensive cost per sample. If an input can be prorated according to yield, then yield maps may be useful for making variable-rate application.

Fertilizers are often recommended based on yield goal. If the yield goal in a field varies, then the fertilizer recommendation should vary. A cropping systems study to determine the spatial stability of yield in corn, cotton and soybeans (and rotations) was established at Alexandria in 1998. Field experiments to apply nitrogen on corn and cotton based on yield maps from a previous season also were initiated.

Steven H. Moore, Professor; Maurice C. Wolcott, Research Associate; and Michael L. Tarpley, former Research Associate, Dean Lee Research Station, Alexandria, La.

(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

Top