In-season scouting is the mother’s milk of any full-service agronomy program. It’s the component of crop production that validates the experience of individuals who know and understand the land, the crops, the climate, and the pests, as well as the interactions of all of them — and who is uniquely equipped to help farmers make best in-season decisions.
“Boots on the ground” scouting will always be essential, but even the best scouting program is limited by how much information that even the most diligent technician can gather and interpret. Tools such as lures and traps and computerized pest models are nothing new and have been evolving for many years. But as wireless telemetry, mobile computing and the miniaturization of everything from sensors to cameras to batteries has evolved, more and better solutions are being released to make scouting programs more robust and better at creating actionable recommendations.
In 2006, the Journal of Agricultural Science based in the United Kingdom estimated that 30% of crops are lost to pest infestation of some measure. Applied to U.S. agricultural production, that’s a $60 billion hit to the agricultural economy. Kim Nicholson, vice president of business development at pest solution provider Spensa Technologies, notes that even a modest 10% improvement in addressing production-robbing pest infestations would return $6 billion in production to the economy – a fact that has driven substantial investment in in-season pest solutions.
“We have lots of interesting tools in agriculture today with traits and technology but the sheer fact is that we are losing yield potential to pest problems,” says Nicholson. “It’s a big battle that we are fighting, even though we as an industry are already spending $15 billion a year in traits and technology. There are a lot of good tools out there, but a lot of the work now is about how we apply those tools, and how we can do it in an effective and efficient way. That is what we spend our days thinking about.”
Spensa is one of a number of companies that are working to combine the best of what we know about crops and pests with emerging technology and field expertise to improve the ability to manage pest outbreaks.
For example, pheromones and pheromone traps have been around for decades, but have required a high level of human interaction to be effective. Existing and emerging trap technology provide a fascinating array of intelligent solutions to attracting, identifying and counting pests that requires precious little human interaction.
Traps across manufacturers have benefitted from improved battery technology and wireless connectivity, so steady power and information transfer, along with pheromone lures, are table stakes. The uniqueness comes from how pests are identified and counted.
Spensa’s Z-Trap employs a proprietary bioimpedance sensor, which can detect distinct species of insects that come in contact with the device. Insects are identified, counted, and recorded by the sensor, and the information sent via cell connection to a database that’s accessible by mobile or desktop devices.
TrapView, a company based in Slovenia, has deployed thousands of its traps across southern Europe in an effort to gain advantage over the cotton bollworm, a significant pest in the region which causes an estimated $5 billion in crop damage annually worldwide, says Matej Stefancic, CEO of TrapView. Developed beginning in 2010 and financed through a grant from the European Union, the TrapView apparatus features a small camera which takes an image of a pest-collecting sticky trap. The image is downloaded and analyzed via computer algorithm to provide a count of pests. The unit is self-cleaning, so the trap surface can be reset without human interaction.
The other major advancement in scouting is the back end gathering and interpretation of data. Spensa incorporates weather data provided through a partnership with aWhere, and brings in scout-generated in-field data and observations via its OpenScout mobile app.
Spensa also offers a dispatch tool that allows an agronomy manager overseeing scouting activity to send personnel out to a field to check on a problem detected by the system.
“The agronomist and the scout’s input is critical, so the tools to facilitate, collect and store information collected by professionals is critical,” says Nicholson. “A solution that is just information is not powerful unless you can use it to make a decision, share that information efficiently, and create value for farmers.”
Armed with an array of data and information, the Spensa system offers tools to help agronomists effectively anticipate and address pest problems. “We have 35 embedded phenology models — either traditional curve or event modules — that help predict when the pest population is at its peak stage for control,” says Nicholson. “We have something called dynamic phenology that incorporates the near real-time trap data with scouting data, which can really narrow down the window for control.”
Spensa is also working on building what Nicholson calls an “automated forecasting network for insect pests,” essentially a weather station for insects, through a grant from the National Science Foundation.
TrapView is looking to deploy in the U.S. through partnerships it is establishing, so service providers can expect to see more of the company in the years ahead. Stefancic says that manufacturers, retailers and cooperatives will be its main partner targets. “The user of our system still needs to be able to understand and interpret the data, so we do not see farmers as our target group,” he says. “Our work in the U.S. this season will be pretty much testing, but then we plan to move fast-forward after that.”
Allan Fetters, Director of Technology for J. R. Simplot, well remembers the travails of scouting fields. “I had to go out and count each individual moth and document the information, then turn around and create graphs and try to understand the windows of control and pest thresholds,” he explained. “So when I saw technology that has the potential to operate autonomously, it was intriguing.”
Fetters has personally been working with in-season tools for about five years, and has used both Spensa and TrapView systems. He wanted to see how those two tools actually perform in the field, so he got involved in testing them early on to see how they would fit. “I wanted to get a sense of the scalability and the potential to reduce the requirement of having summer scouts go out and monitor pest counts.”
Another factor considering Simplot’s diverse crop profile is the increasing introduction of softer chemistries and bioproducts that require more knowledge about the pest life cycle to be efficacious.
“Timing becomes more and more critical if softer materials are being used,” says Fetters. “If you’re on a weekly scouting regimen and a flight of a given pest comes the day after you scout, you’re then six days into an infestation and that will affect what products I can use for control, and probably limit my ability to use a softer product. With real-time or near real-time data coming in, we can maximize our options and be sure we’re applying at the right time and getting full efficacy.”
Fetters sees opportunities in both specialty crops and row crops for high-tech traps. “In some ways corn and some other crops are almost a better fit,” he says. Placing a small number of traps in strategic areas can provide an early warning if, for example, a biotech trait is failing to control pests within a desired threshold. “It can give you an early indication that you might need a spray.”
For high density specialty crops in highly variable land, the opportunity to bring in technology like this is very clear, says Fetters. “I think the technology has a place as we have seen technology costs come down, and is effective at providing the early warning systems that producers desire. Orchards and vineyards have not had the luxury of utilizing a lot of precision agriculture technology – there is a lot of variability from orchard to orchard.” But as effectiveness increases and price decreases, adoption will grow as well.
Sensor As A Service
Out of the Pacific Northwest has emerged a different approach to pest management technology geared specifically to the specialty crop market — setting up a network of hardware and providing the access to the network as a service to the grower and their service providers.
Semios started down this path five years ago on the strength of a $10 million grant from the Canadian government with the mandate to reduce the use of traditional pesticides. The concept: deploy a legion of sensors that offers high resolution pest data throughout key specialty crop growing regions, and attract growers with access to data and targeted sensor deployments within their operations, says Michael Gilbert, Semios Founder and CEO.
Six years into the project and in its third year of commercial operation, Semios has deployed some 150,000 sensors over 35,000 acres. “We offer commercial pest and disease and frost detection, and we are in the early stages of irrigation control,” Gilbert notes. The beauty of the network is that a moisture sensor can be “bolted on” to the existing Semios network, rather than requiring a completely new network be deployed as well.
Specialty crops present unique challenges that the Semios network can rise above. One is the need for a sensor to get deep into the tree or vine where a drone can’t reach. “You need to get it right there in the canopy, and to be effective you need it deployed not every 100 acres, but every acre,” says Gilbert.
Another reason is that every orchard is different, and the impact of sensor use is more dramatic than in row crops. “Apples grown in Washington State might be in an orchard with a 45 degree slope near a lake, or on a hillside,” explains Gilbert, “topography that can change the temperature by 3 or 4 degrees, and the humidity by 10 or 20 percent. Getting granular data on these differences changes decision making, especially when that data is being delivered every 10 minutes on every acre.”
Along with increasing capabilities, Gilbert says Semios is going to continue to work on the robustness of the network with a goal of a half-million sensors. “When it comes to predictive models, one of the most important factors is the size of the data set that will drive the algorithms,” says Gilbert.