By: Ken Cassman, Chief Agronomy Officer, Main Street Data
What is agronomy?
Agronomy is an applied ecological science that seeks to optimize production of food, feed, fiber, and bioenergy through use of crop and soil management practices that are profitable, protect the environment, and maintain or improve soil quality. And although U.S. farmers have made substantial progress in reducing the environmental footprint of the crops they grow, the challenge going forward is to continue this trend as yields rise and management operations become more sophisticated.
How is agronomy changing?
Today’s crop producers must optimize 10-20 crop and soil management practices to stay profitable. Crop rotation, tillage method, varietal selection, seeding rate, row spacing, seed treatment, fertilizer rates and their formulation, timing and placement, weed, insect, and disease management operations, use of cover crops, and whether to use variable rate or uniform application of inputs are decisions a grower makes each year. Classical agronomic science, however, can’t help with optimization of so many interacting factors because it relies on replicated field experiments that can only evaluate two or three factors at once. For example, a typical experiment on corn might evaluate three hybrids, three seeding rates, and three nitrogen fertilizer rates with all other practices held constant. Such an experiment would require 27 to 36 individual treatment plots (9 treatments each replicated three or four times). To obtain useful results, the experiment would need to be conducted over several years at a given location to account for the impact of year-to-year variation in weather, and also at several locations to account for differences in soils and topography. And because crop rotation, tillage method, planting date, and pest management practices all influence which combination of hybrid, seeding and nitrogen rate gives highest yield and profit, the results are of limited value for informing decisions in systems using management practices for “non-treatment” practices that differ from those used in the field experiment.
Hence traditional, replicated field-experiment agronomy is too time-consuming and expensive to provide an effective approach for identifying optimal practices in modern, large-scale, mechanized, high-yield, high-tech production agriculture. But does this signify the end of agronomic science itself? Of course not, but it does mean that agronomic science must change, and the key is to take advantage of the farmer innovation that occurs across the millions of fields planted to crops each year. In essence, each of these fields is an “experiment” that receives a specific set of crop and soil management practices. At issue is how to cost-effectively and efficiently identify which combination of practices works best for a given combination of crop, soil type, and climate.
The power of farmer-reported big data
The good news is that a farmer has company because each of their fields, and management zones within their fields, is likely to have many thousands of “cohort” fields or sub-field zones with similar soil properties and climate such that the same combination of management practices would optimize both production and profit across this cohort group. The challenge is to identify which fields are the appropriate cohorts for a given farmer’s field. Here Main Street Data provides the “Rosetta Stone” through use of 7 years of precise combine yield data over millions of acres and more than a billion subfield zones across the U.S. (Ron LeMay Blog) And those seven years include a severe drought (2012) and years with highly favorable rainfall and temperature (e.g. 2014), which in turn allows development of algorithms that connect each field, and zones within fields, with cohorts that perform similarly over this 7-year period. These cohort fields can be used to benchmark performance of an individual field. For example, if a grower had a field with a yield in the 50th percentile of all cohort fields in a given year, then 50% of the thousands of cohort fields with similar soil and climate had higher yields and 50% had lower yields. Fields or sub-field zones ranking in the 90th percentile tier (i.e. top 10% of all fields in a cohort group) represent the elite performers.
Now imagine a database that contained the key management practices identified above and the GPS coordinates for every production field in all major U.S. crop production regions, and using this database to rapidly identify the combination of crop and soil management practices associated with highest yields and greatest yield stability for each cohort group. The database would be anonymous, secure, and governed by participating farmers who contribute their data in a collaborative, “data cooperative” arrangement. At the end of each growing season, participating growers who share their data could benchmark performance of their fields versus all cohort fields with similar soil and climate and learn which combination of practices works best as indicated by the practices used in the 90th percentile tier fields. Benchmarks could include yield and the associated nitrogen, water (rainfall or irrigation), and energy use efficiencies, which can be estimated by the reported data, in combination with publicly available weather and soil databases. This information would be updated yearly so that the database tracks the innovations driven by grower adoption of new technologies and innovative combinations of management practices.
Farmers’ best interest
The value of farmer-reported data is a “numbers” game. Having access to some of the data is good, but having more data is better because of the ability to identify optimal practices and quantify benchmarks increases massively as the number of observations increases. Balkanization of the data into fiefdoms owned and operated by different large multinational companies is not in the best interest of farmers for two reasons: (i) the power of big data is limited by not having all the data, and (ii) farmers may be limited to technology options produced or supplied by the database owner. Hence the need for a grower-owned and governed data cooperative to house and analyze farmer-reported data for use and benefit of all participating growers.
The Vision is here!
The Growers Information Services Cooperative (GiSC) provides the required platform for the envisioned farmer-reported database because it ensures that contributed data remain anonymous and secure, and the database is owned and governed by participating farmers. (Read Why A Cooperative). A partnership between GiSC and Main Street Data provides the basis for placing each field and sub-field zone in the most appropriate cohort group based on algorithms that utilize climate and soil variables that determine crop response to management. And the power of this approach grows in a virtuous cycle as the number of participating growers increase, and the number of years with different weather patterns expands.
The end is the beginning
So while agronomy is not obsolete, it must change to take advantage of big data and reduce reliance on replicated field experiments to identify optimal agronomic practices for a given field and cropping system. It is in recognizing that each field is part of a much larger cohort group of fields and subfields with similar soils and climate, and that such a group represents a virtual “technology extrapolation domain”, or TED, that the power of big data can be unleashed for the benefit of crop producers to more rapidly innovate and optimize both productivity and profit. And within this framework, agronomists will have a rich feeding ground of high resolution, high-quality crop and soil data to call upon to more effectively and efficiently ply their science.