For most people agriculture research starts and ends with the small-plot work of research scientists and plant breeders at places like the Brandon Research Centre.
Passing by their fields along the Grand Valley Road during the growing season reveals a patchwork quilt of small plots looking at everything from disease resistance to trait expression.
There are some good reasons for the preponderance of small plots: they’re quicker, cheaper and easier than field-scale trials. But they also come with a downside: it can be difficult to extrapolate those findings to larger-scale commercial farms.
That’s why some farm groups and farmers have started to take on issues on a larger scale, something that could be leading to a revolution in the way agriculture research is conducted, and Manitoba is on the cutting edge of this effort.
Five years after setting up an on-farm research network, Manitoba Pulse and Soybean Growers says its by-farmers-for-farmers research is bearing fruit.
MPSG first delved into the practice of putting research efforts on commercial farms in 2010, and launched its full-scale On-Farm Network (OFN) in 2014.
To date the OFN has completed 35 collaborative on-farm trials covering a broad range of topics that delve into agronomic management of pulse and soybean crops.
Megan Bourns, an MPSG agronomist working on the effort, told the recent Manitoba Agronomists’ Conference that the topics were selected because they posed similar challenges across several farm operations — so for example, soybean efficacy trials.
“We use this data to build large datasets that can be used to investigate patterns and probabilities of response across time and space,” Bourns said.
So far that probably sounds a lot like any other agriculture research trial, but here’s where this on-farm research differs: it also tries to delve into questions on the individual farm basis.
“We also look to help growers with questions that are very specific to their operations,” Bourns said. “Data from these trials are meant to stand alone rather than be combined into larger datasets.”
An example of stand-alone effort would be a dry bean tillage system trial where one grower compares several different options.
The three Ps
MPSG’s goal is to test new pulse and soybean products and production practices and there are three key principles: The research must be participatory (conducted on farm, with farmers), precise (produces data that is unbiased, accurate and robust) and proactive (results guide management decisions, improve productivity and profitability of the farm operation).
Trials are developed with observations and input from producers and agronomists to try and answer their questions. Trials to date are addressing 11 different practices or issues such as rolling, row spacing, seeding rate, seed treatments, residue management and fertility in soybeans, and fungicide and inoculant use in pulses.
Research outcomes are shared with individual growers to help them make more informed decisions on their own farm, and are also aggregated to try and come up with recommendations that can be used on a regional or provincial level.
An example is the checklist for single inoculant use in soybeans that came out of OFN research trials. Finally, the data generated helps researchers to guide their efforts.
A trial that has been underway for a couple of years is looking at on-farm soybean rolling, assessing things like sediment movement and surface roughness, to try and quantify the agronomic and economic advantages of soybean rolling on non-stony fields.
On-farm trials are a good example of adaptive science, says Bourns, that is based on sound scientific principles, and provides reliable statistical analysis and meaningful results for producers.
“It’s practical research to answer practical questions,” says Bourns. “With on-farm trials you can adapt your science more than is possible in small plots, and there is more room for creativity as long as the scientific principles remain sound.”
An example would be evaluating strip till with side-banded fertilizer versus conventional till with broadcasted fertilizer in dry beans to compare the systems (fertility and tillage as a package). The trial itself will require replicated treatments covering a large field area.
Field variability is approached differently for on-farm research than it is for traditional small-plot research (where a very uniform site area is desired).
“For on-farm trials, if the research question is being asked at a field scale, then we actually want to encompass as much representative field variability as possible to determine the average effect of the treatment or management practice of interest across that variability,” says Bourns. “If a producer is managing their field in zones, and is looking to apply treatments differently within those zones, then the scale of the on-farm research should reflect that scale of management. In this case, we would want to go ‘within’ the variability and look at treatment comparisons within zones.
“Something we are looking to do more and more is answer the question, ‘why?” says Bourns. “We don’t want to simply determine whether some treatment or agronomic management practice generates a statistically significant yield difference, we want to know what is driving that yield difference.”
A good example is trials that are assessing the agronomic impacts of seed treatment on soybeans, with some trials showing yield response to seed treatments. What’s important is to try and figure out what is driving the yield difference, says Bourns. Were there pests or diseases to begin with in the field and what was the insect and disease pressure during the season, which requires careful and regular scouting of the trial strips.
To get the most out of on-farm trials, says Bourns, it’s necessary to have collaboration of producers and researchers, and make sure it puts the growers and their questions first, asks practical questions, adapts the scientific approach to find practical solutions, and goes deeper into the data collected.
“On-farm trials give the opportunity to be selectively intensive in data collection,” says Bourns.
The power of peers
Jennifer Sabourin, research manager at Antara Agronomy & Research, is a great believer in the power of peer group research.
Antara began its on-farm research network in 2017 because producers wanted more local data to answer their questions, they wanted to use their own equipment and land to know what really works for them, and they were questioning the status quo of some best management practices.
“We were doing a lot of work with commodities groups and the university and noticed that producers who are willing to host these research trials are very progressive, they’re interested and engaged in learning, so they started asking questions about their own farm, their own equipment, and some what-if questions,” says Sabourin. “But, we could not tweak protocols that were being provided to us when we were hired for specific trials, so we thought why not get some producers together who would want to help make their own protocols, and then we can answer questions that they have about their own farm and their own practices.”
The peer group now consists of around 20 Manitoba producers, who each host two trials a year, and who help develop the trial ideas and each host two trials based on their interest and equipment exploring things like seeding rate, row spacing, varieties, fertilizer placement, seed treatments, biologicals and top dressing.
“How we develop our protocols is from the ideas and the questions that our producers have,” says Sabourin. “We pick 10 or 12 trials and those are replicated between at least two, up to five different trial sites based on the producers’ interest, equipment and so on. This year, we had two trials that were hosted on six different sites, so in one season we did six site-years, so we are learning six times faster than if a producer just did one trial on their own farm by themselves. That’s the power of peer group research because we’re choosing together the trials that we’re going to do and then they’re replicated on many sites so that we can learn that much faster.”
Sabourin provides agronomy advice for the trials, is present in the field during implementation and harvest, collects data consistently throughout the season, and provides scouting support for weeds, insects, diseases and harvest timing. With the same person involved at all aspects of the trial, the method and information collected is much more consistent.
“We are taking away the variable of having two or more people involved in the trial,” says Sabourin. “When more people are collecting observations, they might have differing opinions, for example of defoliation by grasshoppers. I am looking at different fields, but I have the same way of judging each plant. I am not adding a bias that you might get with different people assessing an issue. That’s where you can start to get an influence and the data could be somewhat swayed.”
All of the data is presented in an annual research report and includes every treatment strip with all the results and observations explained and shared. The participants have chosen to keep results within the group; nothing is made public.
“It is their data, they paid to participate in the group, therefore it’s their decision and as a group, they decided that the research stays within the group, so it is not put up on the internet, it is not for sale,” says Sabourin. “If anyone else wants access to the research data, they need to participate. If they do, by hosting just two trials, they can receive information and sound, unbiased data from 40 different trials or like trial sites, which is huge.”
A big concern for producers with on-farm research is time, always in short supply at crucial times like seeding and harvesting, so Sabourin says it’s important to help producers be prepared ahead of time so the trials can be as easy and non-disruptive to their operation as possible.
“During the winter, we have meetings, and try to get everything lined up so that the producer knows exactly which equipment is going to be used and when,” Sabourin said. “We think about everything right down to how we can get things done effectively. As an example, if the field is long, we can get a grain cart at one end and then they just harvest on the way back, so we’re collecting variability along the length of the field, yet we still have capacity to weigh it all off.”
The trial location is chosen carefully, to optimize the data that is collected.
“We take a look at trial location in the field, where the hills, valleys, drainage ditches and valleys are, and try to locate the trial where the treatment strips transect its topography, so that the topography affects all the treatments equally,” says Sabourin. “If you run your treatment strips parallel to a ditch, for example, you might have those treatment strips drowned out in a wet year or producing more in a dry year because that is where more moisture would be found.”
Producers participating in the trials often comment that these on-farm trials are the only real place to learn, whether it’s about varieties or practices.
“Our trials are scientifically sound and we can collect incredible information because we’re using such large sample sizes,” she said.