How AI Is Helping Farmers Predict Harvest Dates Amid Extreme Weather

AI-powered crop forecasting tools are helping fruit growers predict harvest dates, estimate yields and plan seasonal labor as extreme weather makes farming more difficult.

Extreme heat disrupts apple harvesting

When the first day of apple harvest arrived in Washington state last year, the fruit was ripe and harvesters were ready. But the weather had other plans.

“It was about 38 degrees… it’s not safe to work in that heat,” recalls Joel Carter of Okanagan Specialty Fruits. “We had to stop at 10 o’clock in the morning.”

AI models that factor in weather conditions and predict ideal harvest dates could help growers, Carter says.

“You need to know more than just when the fruit ripens. How long do you have to harvest it? That’s where these models can be very helpful,” he says.

Okanagan Specialty Fruits owns more than 1,250 acres of apple orchards in Washington. The apples are grown for slicing and are often sold to hotels and schools. The company is investing in technology to maximize productivity. Even some of its apples have been genetically modified so they do not turn brown as easily after being cut.

AI analyzes fruit to forecast harvests and yields

Planning a harvest can be difficult. New tools are emerging that count and analyze fruit on trees and vines to predict when crops will ripen.

This is especially important for high-value fruit such as strawberries and blueberries, where prices can fluctuate widely. If growers get the harvest date wrong, they risk booking seasonal workers they do not need or missing the opportunity to maximize profits.

Okanagan Specialty Fruits is already testing cameras from Canadian company Vivid Machines. Mounted on top of a tractor, the camera captures images of apple trees as the tractor moves through the orchard. AI then identifies buds, flowers and fruit in the footage.

“Right now, Vivid is giving us yield predictions and harvest dates,” Carter says. He adds that the system is particularly effective at detecting very small flower buds that are difficult to see with the naked eye.

However, the accuracy of the predictions depends largely on the quality of the historical data entered into the system, Carter says.

“This is not something where AI can collect information from the internet and figure out an average yield. It’s for Granny Smith. It’s custom-made for your farm,” he explains.

Short harvest windows make forecasting critical

Apples are relatively forgiving. Carter says the harvest period for Granny Smith apples lasts about three weeks. For other fruits, including berries, the window may last only a few days.

“Once the strawberry harvest starts, you have to harvest it, otherwise the whole crop becomes diseased very quickly,” says Raymond Martin, co-founder and chief operating officer of FruitCast, a British company that provides crop forecasting services to fruit growers.

FruitCast provides forecasts for strawberries, raspberries, blackberries, blueberries and tomatoes. “Next year we’ll be moving into grapes,” Martin adds.

Experienced farmers generally know when their fruit will ripen, Martin says. However, it can be difficult to assess entire farms, particularly when they cover many acres or include both outdoor and indoor growing areas.

“We’re doing exactly what farmers can do, but we’re just doing it at a scale they can’t,” Martin says.

The system analyzes footage of ripe fruit captured by drones, smartphones carried by people walking through fields and cameras mounted on farm vehicles.

Extreme weather adds pressure to fruit growers

Martin says the weather has been harsh this year, with large parts of the UK experiencing extreme heat and severe drought.

These conditions stress fruit plants and can cause them to enter thermal dormancy, slowing fruit production. As weather becomes harder to predict, AI crop forecasting could help growers make better decisions about harvest timing, labor and productivity.