Corn+Soybean Digest
Longtime data wizard Gary Wagner Crookston Minn always uses at least two data layers to diagnose agronomic problems Here he compares a yield map to ldquoasappliedrdquo data

Longtime data wizard Gary Wagner, Crookston, Minn., always uses at least two data layers to diagnose agronomic problems. Here, he compares a yield map to “as-applied” data.

Turn agriculture data into field knowledge

Gary Wagner is a 34-year farm-data veteran who’s taken data analysis to the next level to improve his profits. He uses two to four key data layers to identify soil productivity zones and yield patterns. Scrupulous attention to data accuracy ensures meaningful data and conclusions.

Gary Wagner visits his banker with rolling five-year net-profit maps. They document yield stability and yield/profit potential across various weather and crop rotations by management zone.

The self-taught computer programmer and farmer is a sophisticated “big data” row-crop user who has lectured internationally on two major themes: You can’t improve what you don’t measure, and data is worthless unless it’s aggregated in meaningful ways.

Wagner farms 4,600 acres near Crookston, Minn., rotating among corn, soybeans, spring wheat, sugar beets and sunflowers. He defines each management zone by yield maps and remote sensing imagery, plus his own knowledge of the land. Zone size doesn’t matter, as similar productivity metrics drive population, hybrid selection, population and fertility. Some fields have three zones, others 10. He revisits zones every four years.

This example reveals the value of aggregated data versus pretty maps: One soybean field averaged 44 bushels per acre over four soybean years (eight years of a wheat-soybean rotation). Aggregated data revealed that some hillsides never yielded more than 24 bushels per acre, while other areas yielded 55 bushels or better. By targeting his fertilizer based on the data, he raised the field average by 6 bushels while maintaining average fertilizer costs.

Aerial imagery also saves Wagner money on chelated iron, as he can target-apply it to iron deficiency chlorotic (IDC) soybeans or precision-plant defensive IDC varieties. The crop matures more uniformly and is of higher quality at harvest, too.

 

Moisture stress, not compaction

Another example shows the power of perseverance and experience with data: In 1993, using a yield monitor for the first time, Wagner discovered that one field corner was unexpectedly low-yielding over time. At first, he blamed compaction. Four years later, as he dug deeper to identify imagery patterns over the decades – even digitizing old FSA slides back to the 1980s – he began to identify yield dips in the late 1980s, in 1993 and in 1997. Yield monitor data, remote sensing, topography and electrical conductivity (EC) maps over 1,800 acres identified that field corner as having higher elevation, sandier soil and periodic droughtiness. 

“Soil moisture stress caused those yield dips, not compaction,” Wagner says. As it turns out, in 1993 and 1997, the previous crop had been sugar beets, which remove more soil moisture than other crops. The correct solution resulted from not just settling for one technology’s data to solve the problem.

This kind of pattern recognition reveals the true value of big data, Wagner says. Through the Farmers Business Network (FBN), which anonymously aggregates participating FBN neighbors’ data, he can more quickly identify patterns from hybrids, fertility programs and other variables. “For the first time in 22 years, I’m starting to get real value from our yield maps and collected site-specific data,” he says.

 

Garbage in equals garbage out

This raises a vital question: What is the quality of neighbors’ data? Things like faulty yield-monitor data skews patterns and drives false conclusions. “For example, Ag Leader stipulates six yield calibrations for its monitors, two at each of three speeds,” Wagner says.

“One farmer’s yields were lower when he combined at night. He’d only calibrated his yield monitor at one speed, so his night yield figures were off by 10% to 12%, because he traveled more slowly at night. Without the low-flow calibration, his yield values were far less than accurate,” Wagner says.

He believes that farmers who submit their future yield to big-data companies will pay more attention to yield-monitor calibration. “If they don’t, they will be fooling themselves and hurting the entire aggregated community.

“Besides capturing implement-recorded data, real-time recordkeeping is extremely important to our farm,” Wagner stresses. “It’s vital to enter auxiliary information (by tablet or phone) at the time of operation, and not later. This includes tracking every grain load from the field to warehouse, anhydrous ammonia tanks applied, seed lot numbers, grain inventory, crop sales, input purchases and other information necessary to properly interpret the site-specific data we collect.”

John Fulton, precision agriculture expert and assistant professor of agricultural engineering at the Ohio State University, drove the data-integrity point home at a recent big-data conference. “If farm data were cleaned (corrected), 40% of management zones would change,” he says. “Garbage in equals garbage out.”

 

Four key data layers

Wagner banks on these four data layers to drive his bottom-line agronomic decisions, weighing the first two most heavily:

  1. Topography maps drive tile and surface-drainage decisions, “things we can control,” Wagner says. “These reveal drainage problem areas that are the quickest way to make a field more profitable.”
  2. Yield maps are very useful, “because they summarize many soil-productivity characteristics,” as long as your monitor is properly calibrated.
  3. Remote-sensing imagery shows emerging crop patterns. Wagner likes near-infrared (NIR) reflectance imagery as a very reliable indicator of plant vigor as well as cell and canopy structure. His next favorite is NDVI+red (the red shows photosynthesis), followed by NDVI+green (green shows plant health). NDVI, or normalized difference vegetation index, is a ratio of two imageries that indicates emerging vegetative biomass patterns. It’s useful to identify issues such as diseases where canopy vigor is low but previous yields were high, Wagner says.

Imagery is especially helpful if you lack good yield data, because it identifies emerging trends that shape yield, he adds. It particularly helps Wagner with his wheat and sugar beets, as he earns premiums for protein and sugar content.

  1. Bare-soil image or soils map: Wagner gets National Ag Imagery Program imagery from http://datagateway.nrcs.usda.gov/, which sometimes shows bare soils, or from the Web Soil Survey (http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm). These layers are valuable to complete zone delineation.

Other useful data

  • Wagner uses strip trials to evaluate various agronomic practices.
  • A nutrient-removal map (based on yield).
  • Landsat images (medium resolution) are free (http://glovis.usgs.gov/), or $1 per acre or more for high-resolution images. Medium-resolution images are helpful for larger fields, as it’s difficult to manage anything smaller than 100-foot squares, Wagner says. High-resolution images provide greater detail, but they still need to be aggregated into larger areas to be used with today’s application equipment. “Tomorrow might be a different story,” he says.

Imagery interpretation tips

When you look at aerial/satellite imagery or yield maps for agronomic clues, look for irregular shapes. Wagner relies on these principles:

*Irregular shapes come from naturally occurring forces (Mother Nature).

  • Straight lines come from farm management practices (fertilizer/chemical applications, different hybrids)
  •  Patterns or streaks reveal equipment problems.

Weed mapping

Harvest is one ideal time to map weed escapes, Wagner says. Over time, you see patterns useful for prescribing herbicides. He has a device in his cab that geo-locates them with the click of a button. “It’s easy to spot-spray escaped wild oats, thistle, quack grass and resistant weeds once they’re mapped.”

The latest development in profits from data came last year when Wagner had back surgery. He used his down time to scout fields with a new Unmanned Aerial Vehicle (drone). “I could send my brother to the field for ground truthing,” he says.

 

Aggregated data provider

Gary Wagner, a Crookston, Minn., a very early adopter of “big data,” sees tremendous value in aggregated, anonymous farming data from his region. After considerable research, he enrolled with Farmers Business Network (FBN), an agronomic data-analysis network focused on providing independent, unbiased analytics based on anonymous data aggregation.

“It’s probably one of the very few independent data aggregation companies out there not tied to farm input sales,” Wagner says.

FBN “combines your fields’ data with the FBN network’s intelligence to improve your profit potential using powerful data analytics,” FBN’s website says. “Our analytics are backed by billions of data points from other FBN farmers, weather, soil networks, and applied data-science techniques.”

FBN integrates all of your precision data to provide hybrid performance data by soil series and location, yield potential and hybrid matching by field, and yield analysis to see how your practices affect yields.

Founded by Silicon Valley veterans, the service is solely a data-analysis company that helps me identify data patterns to base decisions on. “This is what Big Data is all about,” Wagner says.

 For details, see www.farmersbusinessnetwork.com and http://vimeo.com/120546601.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish