Plots Without Borders: Optimizing Methodology to Conduct On-Farm Research
Term: 3 years
Status: Ongoing
Researcher(s): Steven Shirtliffe, U of S
SaskOilseeds Investment: $120,559
Total Project Cost: $632,170
Funding Partners: WGRF, SPG, Sask Wheat
Objective
The overall objective of this research is to optimize on-farm experimentation to provide precise and accurate agronomic recommendations. We will work with participating commodity groups to optimize the experimental design, implementation, and analysis of field scale, on-farm agronomic experiments.
The end goal is to develop on-farm plot systems that can automatically deploy the trials and analyze the harvest yield maps, so that in the future, data analysis doesn’t need to be done by highly skilled researchers.
Project Description
Engaging farmers in research has the potential to both increase agriculture productivity and reduce the environmental footprint. Farmers’ engagement in research is enhanced by organizing groups of farmers to conduct on-farm experimentation.
This can provide both robust research results and also serve as extension tools to demonstrate the potential value of new agronomic practices. Analysis of these experiments is being facilitated by farmer-centric, digital tools that allow for relatively easy analysis and intuitive interpretation of the results.
Using on-farm trials to guide agronomic decisions on crop inputs is obvious. Compared to small plot trials, on farm trials much more closely represent conditions under which the crop is commercially grown. In on-farm trials, crops are grown using the same equipment and under the exact conditions that occur on farmer’s fields. Furthermore, artifacts that come with small plot trials including differences in crop management, unvegetated spaces between plots and trip line errors are eliminated as the entire area is solid seeded using equipment that is optimized for no-till cropping.
Commodity groups in western Canada have recognized the importance of on-farm trials to demonstrate and evaluate agronomic inputs by dedicating staff to organizing and running replicated on-farm trials, with SaskWheat having the Wheat Wise program and Sask Canola the Top Notch Trial program, SPG has the Pulse Replicated On-Farm Independent Trials (PROFIT) and SaskBarley has BarleyBin Field Lab. As well as other commodity groups also conduct on-farm trials. These trials will be coordinated by WARC this year.
Despite the potential advantages of on-farm trials there are challenges that prevent their use to develop agronomic recommendations. Because of the size of the equipment standard experimental designs such as randomized block designs are difficult if not impossible to use because of the amount of space required. Therefore, strip trials are normally used where one treatment runs the length of the field and the width of the seeder. While strip trials are easy to implement, they do have disadvantages in that the plots are extremely long and narrow and susceptible to differences in spatial variability within the field. Furthermore, they also take up significant space making it difficult to have enough treatments and replicates to describe the curve of a rate response and to give sufficient statistical power to determine differences between treatments. Researchers have proposed shorter strips within the field; however, issues with the sensitivity of grain yield monitors means that the shortest plot length to accurately detect yield changes is over 100m long. As a result, an all-too common result of on-farm strip trials is “no significant difference” between treatments. Unfortunately, this can occur even when there is an agronomic response to the treatment and thus provide incorrect information to farmers. To overcome these issues, researchers have proposed and developed unconventional methods of analysis that can overcome some of the challenges of strip trials including moving window analysis.