Third generation grower and distributer, Paul J Mastronardi (L), CSO & Co-Founder of Heritable Agriculture Davide Sosso and plant scientist Johannes Scharwies (R) at TamBerry Farms in Chatham-Kent, Ont., on April 14.Nicole Osborne/The Globe and Mail
Should you venture into a Toronto grocery store in the depths of winter and purchase a pint of strawberries, chances are the fruit was grown in California and travelled 40 hours from field to aisle.
This system is a natural result of two factors: Consumers want strawberries year-round, and Canada – with a climate subject to cold, dark winters – cannot meet this demand.
But the factors driving the need for a continent-trotting strawberry are being undermined by climate change, which is battering field production, and major advancements and investments in greenhouse technology, which make it possible to grow fruit where no fruit has grown before (at least not in the winter).
However, there is one major hitch in the journey to the next frontier of strawberry production, says Paul J. Mastronardi, a third-generation greenhouse grower and distributor in Chatham, Ont.: For more than a thousand years, the strawberry has evolved to grow outside. Most current strawberry varieties do not adapt well to greenhouses. And developing new fruit varieties can take decades.
At least it used to.
A greenhouse-specific strawberry is the goal of a new partnership between Mr. Mastronardi, Heritable Agriculture (a company founded in an Alphabet innovation lab) and Italian fruit breeding company CIV. And they’ll be using artificial intelligence to achieve it.
“It has huge potential,” Mr. Mastronardi said. “Imagine just plucking [a strawberry] and having it in a grocery store shelf immediately.”
Mr. Mastronardi currently grows strawberries in one of his greenhouses, but the variety was designed for field agriculture and is not high-yielding throughout the winter.Nicole Osborne/The Globe and Mail
It is an example of how AI is promising to advance humankind’s 12,000-year-old system of crop cultivation. However, the fundamental agricultural principles discovered by our ancestors are firmly established, say geneticists. Technology can help, but only the passing of the seasons can determine whether a new crop (and the technology behind it) will bear fruit or wither on the vine.
Strawberries are a volatile product, Mr. Mastronardi says. Production is threatened by climate change, which is driving heatwaves, droughts and rainstorms that decimate harvests, constricting supply and pushing prices higher.
Mr. Mastronardi currently grows strawberries in one of his greenhouses, but the variety was designed for field agriculture and is not high-yielding throughout the winter.
“Our main goal is creating a 12-month program,” he said. “We’re looking to be the everyday supplier.”
Developing a new variety takes around nine years and is expensive, said Federico Stanzani, the managing director of CIV.
CIV, located in the northern Italian province of Ferrara, has been breeding strawberries and apples since the 1980s. When it formed, breeding fruit was mostly a guessing game.
It starts with phenotyping – an assessment of plant traits such as growth, tolerance and yield. Historically, fruit breeders would select plants with desired characteristics and cross them, hoping to eventually reach a product that ticked all the boxes. But success often came down to a bit of luck and took time. It was difficult to meet market demand, Mr. Stanzani said.
The game shifted somewhat in the breeder’s favour around 15 years ago with the introduction of genomics, said Jayasankar Subramanian, a professor of biotechnology, fruit breeding and genomics at the University of Guelph.
Sequencing the genome meant researchers could pinpoint molecular markers – locations within the genome that coded for specific characteristics – and track changes on them after breeding.
However, genomics drove a “data deluge” – sifting through it took months, and finding patterns was nearly impossible.
For more than a thousand years, the strawberry has evolved to grow outside. Most current strawberry varieties do not adapt well to greenhouses. And developing new fruit varieties can take decades.Nicole Osborne/The Globe and Mail
AI can dismantle that deluge, said Brad Zamft, the chief executive officer of Heritable Agriculture. It can predict the value of complex genetic traits and find genetic combinations that will make a strawberry suited to greenhouse growing.
The first batch of strawberries created by the partnership arrived at Mr. Mastronardi’s farm in mid-April, alongside another location in the province.
They were trial plants. CIV provided a collection of diverse strawberry varieties. When they are grown, the desired properties will be measured. Heritable’s AI will then be able to analyze that data and pinpoint what needs to be adjusted.
They hope to develop a new variety within five years – hopefully, three.
It should be a lucrative venture, Mr. Zamft said. Strawberries are a premium product, and Ontario’s greenhouse sector has already proven itself.
Over the past decade, the sector has grown to cover 4,000 acres and become the second-most concentrated in the world, according to a report from CIBC. It has contributed $6-billion to the province’s economy and, as of 2023, Canada’s greenhouse sector (dominated by Ontario) accounts for 48 per cent of the value of fresh produce exports, according to Agriculture and Agri-Food Canada.
The advantage of the greenhouse is control over inputs. Light, humidity and temperature can all be set to an ideal level. A Canadian greenhouse operation yields 8.5 times more produce per area of land than traditional field agriculture, and greenhouse varieties yield year-round.
Genomics promised to radically transform the way fruits are cultivated. While phenotyping is more informed today than it was 40 years ago, it is still a hard science.Nicole Osborne/The Globe and Mail
AI has great potential for indoor fruit production, said Prof. Subramanian. In this way, it is akin to genomics. Knowing the genetic makeup of a peach tree – Prof. Subramanian’s specialty – helps him develop a new variety in about 15 years, compared with the 20 it took before the technology.
But he warns against too much hype.
Genomics also promised to radically transform the way fruits are cultivated. While phenotyping is more informed today than it was 40 years ago, it is still a hard science. Breeders might select for certain characteristics but miss other molecular markers important to the crop’s health or viability, and this shortfall shows up in the field or the greenhouse.
Ultimately, each new variety still needs to survive the passing seasons – regardless of the growing conditions. While more protected from extreme weather and pests, light levels and temperatures in a greenhouse change throughout the year, Prof. Subramanian said.
Nevertheless, the technology is promising and – rather ironically – might allow farmers to return to more traditional values when it comes to growing food.
For decades, crop cultivation has focused on yield, Prof. Subramanian said. Breeders have selected for resistance and high productivity, and this has often come at the cost of taste. Watery, flavourless tomatoes are a common example.
AI, through its ability to find patterns and sift through mass amounts of data, is able to co-optimize across the board, Mr. Zamft said. This means future fruit varieties will be better able to balance the growing demand for food with consumers’ appetite for taste, he said.
“We believe we can do more. It’s not a zero-sum game.”