2.1 Why is yield monitor data quality important?

Georeferenced yield monitor data is considered one of the most valuable data sets collected during the growing season.

To many people, spatial yield data is the most valuable data that is gathered each year.

Yield data can be used to develop management zones, develop fertilizer recommendations, assess hybrid performance, evaluate the performance of a product or practice through on-farm research, and can be used to assess spatial field profitability.

However, there are errors that are inherent in nearly every sensor or system of sensors used in precision agriculture – including yield monitor sensors.

Many of these erroneous data points can be removed before using the data further. This is often referred to as cleaning yield monitor data.

Let’s take a look at an example of what can happen if erroneous data points are not removed.

Figure 1. Raw yield data zoomed in to the end rows of the field.

In Figure 1 you see a raw yield data map zoomed in to the end rows of the field. The red data points are areas of low yield or no yield, whereas the green data points represent higher yield areas. Notice how low yielding, red, data points are present where the system continued to record yield while the combine turned at the end of the rows and drove over areas that had already been harvested. In these areas, a yield of 0 bu/ac is being recorded. This is an example of a header position sensor error. We will talk more about different types of errors in the next video.

If we used this corn yield map to determine yield goal in order to create a spatial nitrogen fertilizer prescription, this is what the prescription would look like (Figure 2).

Figure 2. Raw yield data (left) and nitrogen fertilizer prescription created using raw yield data (right).

In this case we were used the University of Nebraska – Lincoln nitrogen fertilizer for corn recommendation to develop the prescription. The errors that were in the yield monitor file have affected the prescription that we generated.

Now let’s take a look at that same yield data map, but this time we have cleaned the yield data (Figure 3). Here you can see that areas where the combine was turning and no grain was being harvested have been removed. Adjustments have also been made to adjust for flow delay.

Figure 3. Yield data map which has been post-processed or cleaned using Yield Editor software.

The before and after yield map is in Figure 4.

Figure 4. Raw yield data map (left) and cleaned/post-processed yield data map (right).

Now we can use the cleaned yield map to determine our yield goal and generate a spatial nitrogen fertilizer prescription.

If we compare the nitrogen fertilizer prescription we generated using raw yield data to the nitrogen fertilizer prescription we generated using clean yield data, we can see that it looks very different (Figure 5). In the fertilizer prescription generated with the raw yield data, the nitrogen fertilizer rate was improperly lowered. In this scenario, certain areas of the field would have received 75 pounds per acre lower nitrogen than was recommended with the clean yield data.

Figure 5. Nitrogen fertilizer prescription which was generated using raw yield data (left) and cleaned yield data (right).

If some type of yield data post-processing or cleaning had not been completed, errors like these could affect future management. In this case, the lowered nitrogen rates could result in lower yield in these areas for the next crop. In this way, the errors would continue propagate.

This example demonstrates the importance of data quality when looking to make use of the yield data that is being collected. In the next lecture we will look at some common sources of error in yield monitor data.

Complete and Continue