Drone pollination of apples has now been tested in the orchards of New York state. According to a tweet from Dropcopter, the company experimenting with the service, the “first ever autonomous pollination of apples” occurred in the middle of May. By early June, local media reported that a hexacopter drone had been used to dispense pollen on five acres of apple trees at Beak & Skiff Apple Orchards in LaFayette, New York. The “Worker-Bee” drone flew back-and-forth over the trees at around eight feet and sprayed the pollen from a device on its underside.
In an interview with local radio, Adam Fine, founder and CTO of Dropcopter, said aerial pollination on this scale has never been tried before. “We are actually the first in the world to try it in a real world setting,” he said.
But apples are not effectively pollinated by wind. They depend on insects or other animals to physically transfer pollen from the sexual organs of one flower to those of another. (Other proposals for pollination by drone, like Walmart’s recent patent, are based on replicating this physical transfer.) So the question is: How effective can aerial pollination actually be? Especially in comparison with bees.
“I have no idea,” says Peter Fleckenstein, a partner and the director of fresh fruit operations at Beak & Skiff. “I’ve never even thought of this before.”
According to the Dropcopter website, previous tests conducted in almond orchards between 2015 to 2017 resulted in “a dramatic crop set increase of 10%”. The goal now is to determine whether or not this kind of drone pollination can also increase yields in apples. This is exactly why Fleckenstein agreed to let Dropcopter run a trial in the orchard.
“We’ve got a ton of healthy bees here,” he says. “We have very, very strong hives and good bee work.” However, in the Northeast, the warm weather that brings apple trees into full bloom can be followed by weather that is cold, rainy and cloudy. And honey bees don’t work well under those conditions. Being able to call a drone service to provide pollination when such circumstances arise would be extremely helpful to apple growers. “My interest was in mitigating the weather factor in the pollination,” he says. “Because if you can take that out, that’s just one less thing you have to worry about.”
So how will we know if Dropcopter’s approach actually results in increased fruit set among the apple trees at Beak & Skiff? How can we be confident that any increase isn’t the result of some other factor?
Dropcopter declined a request to address these questions or provide comment for this story, saying the company is not currently providing interviews. Dropcopter also has not made public any data or information about its tests with almonds, so we can’t determine the accuracy of its claims of previous performance or get a sense of how the company might be structuring this particular test with apples.
However, Fleckenstein did explain that the orchard block that Dropcopter used was made up of rows that were equal in length, had trees that were all the same age, and only contained two varieties of apples. There were eight control rows and eight treatment rows. He says Dropcopter hired a research firm to collect data and analyze differences in fruit set.
“The challenge we have this year is obviously there were honey bees present when they were flying,” he adds. Which means that, at best, the results of this recent test will only reflect how the drone performed in addition to the bees already working the trees, not all on its own. At worst, the results might not give us any real indication at all of how well aerial pollination works in apples. Was honey bee (and wild bee) activity comparable in both control and experimental rows? And what other factors might have come into play? How were all these things assessed during the test and how will they be accounted for in the results?
But even with what might be less-than-ideal controls for the experiment, Fleckenstein would feel confident saying that a significant increase in fruit set could be attributed to the aerial pollination and not some other natural factor. He says that within the same orchard block, in the same geographic location, among the same variety of apples, he typically sees very little natural variation.
However, the key consideration for this year’s results is how quickly Dropcopter’s data analysis team got into the rows and assessed fruit set.
Fleckenstein says growers only need 25 to 35 percent of the blossoms on a tree to become apples; this ensures the tree will put its energy and resources into producing fruit that are the right size, shape and taste. That being said, growers do want every flower on the tree to get pollinated so they can go through later and manually adjust their crop loads to the preferred amount. This means Dropcopter’s data team had to get its measurements before the Beak & Skiff crew started knocking nascent fruit to the ground — which the crew is now in the process of doing.
“The key will be: Did they have the research people right there with them the whole time?” says Fleckenstein. “They can’t just come in at harvest and say we got more bushels out of this row than that row. Because we’ve been messing with it now.”
Fleckenstein does not know if or when Dropcopter’s team collected data this year.
When asked about the contribution that wild bees make in his orchards, Fleckenstein says he doesn’t have a good sense of how much these species might be moving pollen from one flower to another. He says that researchers from Cornell University have previously sampled the native bees at Beak & Skiff but that he wasn’t aware of what the results were. And even though some people might feel that apple growers don’t need to bring in honey bees because they can rely on wild species instead, it’s hard to take that risk as a farmer and business owner.
It’s currently unclear if, when and the extent to which Dropcopter will share the results from this year’s experiment.