Data offers more predictability for management decisions

The Ohio State Precision Ag team continues to tally up megabytes as the effort to gather a world record setting amount of data for a single corn plant (named Terra Byte). The project is also seeking out which data is most valuable for making agronomic decisions.

The effort includes a partnership with Integrated Ag Services (IAS) that develops seeding and fertility management zones called common production units (CPUs). CPUs were created to help make farming more predictable and, therefore, more profitable.

The CPU process starts with the IAS automated precision soil sampler that slices nearly seven inches deep for 30 feet in length, allowing for a soil sample that represents the entire soil profile. The soil samples are placed in cups where they receive a QR code and then sent to the lab. The automated sampler designed by IAS collects samples across fields much faster than standard testing and can complete the task in wider range of soil conditions. For the Farm Science Review field that is home to Terra, the fields are sampled on a one-acre grid in the fall every four to five years.

“So Terra started out with a planting prescription that was derived from our CPUs that are based on intensive soil sampling, data collection, and normalized yields that come together to develop a map for the entire field,” said Andrew Hofner, an agronomic consultant with IAS. “We use organic matter and the cation exchange capacity from the soil samples to accurately map yield limiting factors. Slope and topography data are also included — flat soils have better yield potential than sloped. There is also a historical yield layer that reflects more information that we may not pick up in the soil data. That will really show your drainage and other production factors.”

Once the data is cleaned and normalized, numerical CPU designations are applied to different parts of the field. The IAS CPU scale is 1 though 9, with 9 being the highest productivity potential.

“This designation allows us to standardize yield potential across multiple environments. A 5 in Madison County now represents a 5 in Champaign County. What this characterization allows the grower to do is delineate higher yielding areas in fields versus lower yielding areas to make site-specific decisions. We map out areas that are consistently high yielding versus consistently lower yielding,” Hofner said. “The field in question at the Farm Science Review contains CPUs ranging from 5 to 9 — it contains mostly Kokomo and Crosby soils. Sometimes you’ll see much more variability than that.”

In terms of seeding rate, for example, the range for the field was 32,000 to 36,000 seeds per acre. The 5s were in the 32,000-population range and the 9s were in the 36,000-population range.

“The recommendations vary with the grower’s propensity for risk. It comes down to what the grower is comfortable with for the final decisions,” Hofner said. “Then after harvest, we use the CPU as a way to judge agronomic and profit decisions.”

The CPUs do not necessarily account for a huge portion of the total bytes collected for Terra, but they are among the most valuable data in terms of agronomic decisions and return on investment, said Nate Douridas, the farm manager and CCA at the Farm Science Review.

“Right now we are working on seeding rates and prescriptions for nitrogen but moving forward, I think there is a lot of potential. Any time we need to go out and quantify something in the field — whether it’s an action item we have taken to keep the crop going or we are looking at an opportunity to improve something, maybe it’s a fungicide application this year, revised drainage plan for next year or a hybrid change moving forward — the first thing we go to now are the CPU maps. We can set apart the different zones in the field and we can do that with a lot of confidence,” Douridas said. “We’ve found this as the best way to accurately designate our use of seed prescriptions and variable rate nitrogen on the field and even impact actions like scouting. Before we had this CPU map, we didn’t have a whole lot of confidence about where those spatial differences were across the field. We had the soil survey maps from years past to work with, but we know that in some cases those can be very far off from what is actually in the field.”

Douridas said the CPUs offer much more reliability than soil type maps.

“These CPUs clearly have differences from the soil type maps. These are the high definition zones we use for management now. The imagery we take in-season is an additional correlation,” he said. “IAS does the seeding recommendations then we look at the hybrid itself and the weather and make in-field decisions on seeding rates at planting whether to use the seeding rates recommended or a little higher or lower. Then, I can also find out how a hybrid did beyond my farm and performed against the IAS pool of data. How did this hybrid do on other 5s from around the state compared to the 5 CPUs in our field?”

Graduate student Trey Colley stands with Terra Byte in the field at the Farm Science Review.

Graduate student Trey Colley stands with Terra Byte in the field at the Farm Science Review.

The field where Terra is growing is one of the better performing fields for the Farm Science Review. Terra is in a CPU of 8 with an emerged stand count of 34,451 with around 35,000 seeds dropped. There were 180 total pounds of nitrogen applied with 40 units applied at planting on May 16 and the balance applied with a June 9 sidedress application.

By early July, the Terra data collection effort had just passed a billion bytes and the number has climbed from there.

“That adds up to 4.3 gigs of data just for Terra and if we multiply that across the whole field we are talking 13.8 petabytes,” said Trey Colley, a graduate student in precision agriculture at The Ohio State University who is leading the effort. “There are a lot of zeros involved.”

Colley said the CPUs account for roughly 10% of the total data collected so far, but it offers a big bang for the buck.

“That comes back to the conversation of what types of data is actionable and how can we make informed agronomical and business decisions from these data types,” Colley said. “What it comes down to is what’s most economical for the producer.”

 

To follow Terra’s progress through 2017, follow @OhioStatePA on Twitter and Facebook for weekly updates. This is the second story of a four-part series following the project through 2017.

Print Friendly

Related Posts

Leave a Reply