Farm Progress

Extension Crop Connection: Think about precision-ag data management like a Rubik's with six sides to address — collect, store, clean, analyze, prescribe and verify.

November 2, 2017

4 Min Read
MULTIPLE SIDES: Collecting yield data is the first step in data management. But even if yield monitors are well-calibrated, errors can still occur. That's why post-processing of data matters.

By Nathan Mueller

Collect, analyze and prescribe. Is precision ag that easy? Let's equate precision-ag data management to a Rubik's Cube. You have six sides to address, not three: collect, store, clean, analyze, prescribe, and verify.

Transforming an array of jumbled colors into a successfully color-coordinated cube isn't easy. We want to create actionable information with precision-ag data that will help us improve economic viability, address spatially variability, reduce environmental risks and boost resource allocation.

So what sides of the cube are stopping you from leveraging your on-farm data?

Stuck on one side: Clean
Most realize that the data we collect includes errors. We assume that a few errors or bad data just don't make that big of a difference when averaged with lots of good data. Yield data is one prime example. We did a great job of calibrating our yield monitor and that is enough, right? Unfortunately errors in yield data still exist after calibration due to velocity changes, header cut width and overlap, flow delay, and more. Yield data is used to develop fertilizer recommendations, create management zones, assess hybrid and variety performance, evaluate product performance, and assess profitability. Do we want to include this bad data when conducting analysis and making prescriptions?

We can remove many of these common errors in our yield data through post-processing or cleaning. However, addressing a suite of errors found in yield data is not a part of many precision-ag software and services offered. The impact of not cleaning yield data could make a 10-bushel-per-acre difference in soybean yield for a particular 60-by-60-foot grid, it could make a 15-pound nitrogen recommendation difference, or could lead you to assume an area of the field didn't yield well, when in fact it did.

A software program called USDA Yield Editor 2.0.7 can be downloaded and used for free. The USDA Yield editor has both automated and manual cleaning options. We can create a cleaning template for a combine, operator and crop. It has been well worth my time to use the USDA Yield Editor to clean my own yield data and on-farm research data for farmers I work with.

You get more information on cleaning yield data in our recent publication, “Improving Yield Map Quality by Reducing Errors Through Yield Data File Post-Processing” at extensionpubs.unl.edu.

Solve multiple sides: analyze and prescribe
I think everyone has a part of a field that just doesn't yield well and you don't know why. In many cases, growers ask professional agronomists to help in these situations. As one of those agronomists, I have concluded that soils can hold a grudge from mistreatment for a long time. One of the tutorials in our past year's UNL precision ag data management workshops was to show how we can use old aerial imagery from the 1930s to 1970s to explain current spatial yield patterns. We can get old aerial imagery from several sources and then georeference the imagery using precision-ag software.

As more land is rented from absentee landlords, we have lost information about the farm and field history. For example, 30 acres of this 160-acre farm was in pasture until in the 1970s or that field used to be farmed east-west as four small 20 acre fields, but we now it is farmed north-south as one 80-acre field. Both of these are examples that have explained spatial variability in yield maps of today.

Finishing last side: verify
The last side of the Rubik's Cube to solve is to verify prescriptions through on-farm research. It is important that when you make variable-rate seeding, multi-hybrid planting, variable-rate nitrogen, management zone and other prescriptions that you build in replicated checks to verify that your prescription was effective (yield improvement, cost saving, etc.). Far too often we make a variable-rate prescription without including verifying that it was effective and accomplished the goal. The Nebraska On-Farm Research Network helps farmers conduct research on farm, in their field and with their equipment. Farmers through the network have evaluated the effectiveness of their prescriptions. You can view grower results through the new On-Farm Research Network Database at resultsfinder.unl.edu.

Solve and resolve
Just like solving a Rubik's Cube, leveraging precision-ag data isn't easy. With patience and persistence you can solve and resolve all six sides (collect, store, clean, analyze, prescribe, and verify) as precision-ag data management continues to change.

Mueller is a Nebraska Extension agronomist in Dodge and Washington counties.

Subscribe to receive top agriculture news
Be informed daily with these free e-newsletters

You May Also Like