PPDL Picture of the Week for

January 26, 2015

Wandering Swath Width Syndrome

RL (Bob) Nielsen, Extension Agronomist

This is not your typical crop diagnostic problem, but one that can nevertheless result in incorrect estimates of crop yield and, consequently, incorrect agronomic interpretations. While we often focus on the importance of yield monitor calibration relative to logging accurate yield estimates during grain harvest, there are other yield monitor settings that can inadvertently influence yield estimates. One of these is the option in certain displays to automatically adjust harvest header or swath width based on the harvested "coverage map" and the estimated current geo-position of the combine in the field. Header or swath width, of course, is used by the yield monitor to estimate the harvested area and the calculation of yield per acre for individual data points and so accurate widths are important to ensure accurate yield estimates.

When set to automatically adjust header width, the yield monitor will automatically decrease the header width if it perceives that part of the combine header is overlapping a previously harvested area. When the estimated geo-position of the combine is accurate, this automatic setting is great when harvesting point rows or field edges in corn or when harvesting soybeans in general. However, when the estimated geo-position of the combine is not accurate, the yield monitor may erroneously change header widths in the middle of the field where, in fact, the combine is NOT overlapping a harvested area.. This is most likely to occur when the combine is using DGPS signals from WAAS or similar signal sources with positional accuracies ranging from 5 to 15 feet horizontally. 

The consequence of inadvertent and erroneous header width changes on yield estimates can be quite significant. Figure 1 illustrates a small section of a field where the yield monitor erroneously decreased the header width value from the full 20 feet (eight 30-inch rows of corn) to 17.5 feet (seven 30-inch rows) for a short distance. 

The incorrect and narrower header width values for those data points resulted in overestimated yields per acre for those data points because the estimated harvested area for those points was erroneously smaller. Figure 2 illustrates the estimated yields for the individual data points. The average estimated yield for the 8 data points with incorrect header width values was 263 bu/ac. The average estimated yield for the surrounding data points with the correct 20ft header width values was 223 bu/ac. The data points were logged every second at an average speed of 5.7 mph and so are approximately 8 feet apart, meaning that there is approximately 64 feet of incorrect header width values and, subsequently, incorrect yield data. 

See my lengthier online article for more information about Wandering Swath Width Syndrome:

http://www.kingcorn.org/news/timeless/AutoHeaderWidth.html

​Click ​images to enlarge

1 tn
Fig 1: Yield data points depicting full (20ft) 
header widths (blue) and incorrect partial 
17.5ft header widths (yellow).
2 tn 
Fig 2: Yield data points depicting estimated 
yields. Average yield for the 17.5ft header 
width points was 263 bu/ac while the average 
yield for the surrounding normal 20ft header 
widths was 223 bu/ac.​