"New problems require new solutions." For University of Guelph professor Dr. Asim Biswas, the world is full of new problems.
By 2030, the global population will reach 8.5 billion, driving an immense demand for food. At the same time, climate change is weakening food systems by depleting soil, a resource so precious it takes 100 years to form a single inch.
Dr. Asim Biswas
As the School of Environmental Sciences professor writes in Nature, there is an "urgent need for improved measurement, reporting and verification of soil health."
For World Soil Day 2024, he says data is the answer - cheap, rapid and accurate soil data - and that his research into affordable technology can transform the way we feed the world.
"Our goal is a seamless integration of technology, data and decision-making," says Biswas, the Ontario Agricultural College Chair in Soils and Precision Agriculture. "We cannot stay with the old system."
Out with the old, in with the instant
Today, the same "old system" of soil health management relies on lab testing that takes days to deliver results - time farmers can't spare.
"If I get a pH reading of 7.55 from a lab, but 7.6 from a sensor, both require the same resources to fix," Biswas says. "In that context, the data may not be 100 per cent correct, but the trade-off is you can gather a larger amount of data more inexpensively."
Biswas, recently named the Canada Research Chair in Digital Agriculture, says his core research vision is to integrate handheld detectors, open-source platforms, digital sensors and artificial intelligence to give similar data in mere moments.
In short, the tools and information are out there, he says, but the goal is to build new connections among them, giving back the time, energy and resources that might just be needed to secure the world's food supply.
In practice, that vision means digital sensors scattered across our fields, implanted in the ground or mounted on tractors, aerial vehicles and remote satellites. These sensors report from the ground or fly above to observe farmers' lands, creating soil data nodes that can be synthesized into a map.
Such rich data help farmers make critical choices to improve their use of fertilizer and pesticides, which could, in turn, reduce greenhouse gas emissions, conserve water and increase income.
The spectrometer devices he's designed also work right in farmers' hands. These devices shoot various types of light, or electromagnetic signals, to obtain information on the soil's nutrients. How the light returns to the device - its changes in intensity, colour and energy, for example - creates a "soil signature," instant data on organic matter, nutrients, pH, moisture content and more.
His spectroscopic device, SoilReader, is already being used by Canadian producers. His data-sharing solution, SoilPrint, keeps farmers' information secure and private by anonymizing key details.
"We're aiming for very low cost, low time commitment and labour requirement solutions," Biswas says. "That is how we feed the future."
Smarter soil breaks through AI ground
The data can then move to the world of artificial intelligence. Machine-learning algorithms can generate high-resolution soil carbon maps - digital field guides that tell farmers how nutrients are distributed and where to adjust.
Soil samples in the laboratory of Dr. Biswas
These techniques, developed by Biswas and already being used in the industry, are building a new era of precision agriculture.
"We have to bring solutions to the farmer that speak in a language they understand," Biswas says. "If we just spit out a pretty graph, that won't help them. But if it tells a location - go to this corner of the farm and apply 60 kilograms of nitrogen - that will connect to the farmer."
Speaking the language of farmers drives his passion project, SoilGPT, a large language model similar to ChatGPT but tailored to provide detailed answers on Canadian soil.
Harvesting data from open-source locations - government websites, university research and libraries - it aims to make a one-stop resource for Canadian farmers. Soon, voice-generative models will let farmers speak directly to SoilGPT while working, creating a hands-free AI companion on the field.
Though Biswas jokes SoilGPT might complete his undergraduate students' assignments for them, he says we can't shy away from the technology.
"AI is developing in every sector, and we have to embrace it."
Still, convincing farmers to move away from traditional methods remains one of the largest challenges. As urban development threatens high-quality soil, solutions to make current agricultural land more efficient can't come any sooner.
"If I'm a farmer who's worked for 40 years, why would I adopt a new technology now?" Biswas says. "Somebody needs to give them the confidence that these technologies will be more effective than the status quo. That is what motivates us."
