Written by Jonathan Eisenthal
Surveys show two-thirds of crop farmers now use some form of precision ag technology. Data collecting/crunching software helps farmers make the most of global positioning systems (GPS), auto steer and variable rate planters and sprayers.
The 2015 University of Minnesota Production Agriculture Symposium drew some of the top experts in the science of lining up management decisions with Big Data. Mapping soils and how crops respond to them is the current star of the Big Data revolution on the farm.
But hardware still gets a lot of attention, too. One speaker, Todd Colten of Minneapolis-based Sentera Corporation, conveyed the latest trends in Unmanned Aeronautic Vehicles (UAVs), more popularly known as drones. Whether configured as airplanes or helicopters, these flying machines become a kind of eye in the sky, gathering data about fields and offering day-by-day field conditions throughout the year.
A small group of exhibitors included a new company called Rowbot. This past year, the firm offered the first robotic management of corn in Minnesota, according to company representative Chris Reedy. The company comes out to the farm like any custom operator, but Rowbot brings a fleet of small machines that run up and down rows of high corn — stage V-6 and higher — dispensing nitrogen as efficiently as any high-boy. On-board sensors, linked to GPS, keep the machines in the row. They also offer cover crop seeding.
But Big Data was the order of the day. Several of the speakers made mention of the major, commercially-available data-crunching software systems: Adapt-N, Climate, Encirca and Maize-N. These systems take the data and model the likely outcome.
“There are a lot of big ideas that aren’t quite ready to make farmers money,” said Karl Nesse, now in his third year as a crop consultant for Centrol Crop Consulting, a firm based in Marshall. He believes that more proof needs to come, through independent duplication of results.
“At the same time, there are a lot of current practices that are not universally adopted that are making farmers big money, such as variable-rate seeding, variable-rate fertilizer, electrical-conductivity mapping, yield data analysis,” Nesse said.
Two of the biggest names in “big data” on the farm, Dr. Rajiv Khosla and Dr. Brent Myers, gave talks about this incipient Ag Information Age.
Dr. Khosla is the coordinator of a degree concentration program, brand new at Colorado State University, called “Applied Information Technology in Agriculture.” He also runs an outreach program that helps Colorado farmers adopt precision agriculture techniques, and modify them to suit their needs.
Khosla summed up his message by saying, “Embrace variability.”
Farming by averages should be history, he said. Historically, without any other way to do it, farmers have had to come up with an “average” level of nutrients and chemicals to apply across fields and even whole farms. This practice puts the producer in the wrong 90 percent of the time — either over- or under-applying inputs.
Digital yield maps and grid soil sampling at a much higher resolution can help farmers properly adjust their inputs.
Myers was part of the team that developed a precision modeling application called Encirca. He told the group of professors, students, farmers and ag professionals at the symposium that ’Big Data’ really began 100 years ago, with the launch of the National Cooperative Soil Survey. Shortly after 1900, experts literally walked the farm fields of America and collected soil data of every acre of farmland in the country. The result: SSURGO, the Soil Survey Geographic database. It is both maps and data in tables. NRCS still oversees the online version of this information.
“Don’t throw out your traditional soil maps,” said Myers. “They contain very important information we need. But maps at a scale of 1:24,000 mean the smallest increment is four acres. This doesn’t match the scale of today’s agriculture. Compare this to digital yield maps where you might have 30,000 data points.”
Up to a point, the more soil samples, the better.
Khosla likened soil sampling to tossing pebbles into a pool. Just as pebbles that hit too close together create interference patterns in each other’s ripples, collecting data from spots too close together can actually create a false picture of soil types and other features of the landscape. It’s a fallacy called ‘spacial dependence.’
Mathematically, he and other scientists arrived at an optimal spacing of samples at 7.5 meters apart, still a far higher sample rate than the typical grid sampling most farmers use today (2.5 acre increments). Khosla said farmer experience and other data can be used to formulate a system of ‘smart sampling.’
“There’s nothing wrong with grid soil sampling, if we can do it at the right scale,” said Khosla. “Smart sampling means finding homogenous zones and doing composites from those.”
Precision ag management software systems are beginning to point the way to major savings and improvements, with more to come in the future, according to Myers.
Variations in weather have made it a prudent management practice to allow for nitrogen loss.
Myers said, “Adapt-N, for example, is a program developed by Cornell University. Adapt-N showed the difference between fall application of nitrogen and split application: it showed how, in test cases, the crop got 45 pounds of the 180 pounds of nitrogen delivered with a fall application, alone. Splitting that same amount of fertilizer, fall and spring, the crop got 95 pounds of the 180 pounds used.”
In addition to managing the farm in a timelier way, precision ag can help the farmer understand space differently, and so improve management. Variable rate application, geared by GPS to the specific features of each field, promises to deliver nutrients and chemicals as needed. More in some places, less in others.
“With the ‘smart sampling method’ (SSM) we can maintain the same level of inputs, just distributed differently, and get higher yield,” said Khosla. “In some cases we have cut back inputs and gotten the same yield as the historical average for the field. We’ve seen cases where higher inputs, applied judiciously, get higher yields. The best scenario for precision ag management is to use fewer inputs, but achieve a higher yield. Of course, the worst scenario here is to cut inputs and get lower yield. However, in our work, over the past 16 years — that’s 70 site years— we’ve had only one case of getting lower yield from SSM.”
The 2nd Annual University of Minnesota Production Agriculture Symposium was organized by graduate students at the university’s ag campus, in the departments of Horticulture; Agronomy and Plant Genetics; Water, Soil, and Climate; Plant Pathology; and Ecology, Evolution and Behavior. Monsanto provided funding for the event.