Soybeans growing next to 2 Minnesota solar farms drew twice as many bee visits as soybeans farther off, and the native bees doing the work had multiplied twentyfold under the panels
There is a fence, and there is a soybean field on the other side of it.
Inside the fence, rows of dark panels stand over waist high prairie in flower.
Outside, an ordinary crop, planted by somebody who had no say in any of it.
Seeded plants spend their early years putting roots down rather than flowers up, and the first counts under those panels found sparse bloom and very few insects.
The bees do not know about the fence.
Counting them turned out to be the interesting part, because they were not staying put, and the farmer was getting something for nothing.
Where did they all come from?
Why the ground under a panel stops behaving like a field
A raised solar array changes two things about the dirt underneath it, and both of them favor prairie.
The first is heat. Panels take the hardest midday sun off the soil, and the plants below get a cooler, damper root zone than open ground offers in a Midwest August.
The second is water. Rain does not land evenly any more. It runs off the panel surface and arrives in a concentrated line along the drip edge, which leaves the ground under an array patchy by design, wet in strips and dry between them.
That patchiness is a problem for a crop, which needs every row alike.
For a prairie mix it is closer to home conditions. These are plants shaped by uneven ground, and dozens of species sorting themselves into the wet lines and the dry ones is exactly what a seeded prairie is supposed to do.
What was planted, and how slowly it worked
The two sites sit on retired farmland in southern Minnesota. Before the first summer of operation, the soil under and around the arrays was seeded with dozens of species of native grasses and wildflowers rather than the usual gravel or mown turf.
Then almost nothing happened.
Prairie is slow. Seeded plants spend their early years putting roots down rather than flowers up, and the first counts under those panels found sparse bloom and very few insects.
The change arrived on the plants’ schedule, not the operator’s. By the third growing season the flowering had thickened enough that the insect numbers stopped wobbling and started climbing, and they kept climbing until the counting stopped.
Four fifths of all the native bees seen across the whole project were seen after year two.
What five years of counting produced
The work was deliberately unglamorous. Walk the same transects, four times each summer through the peak flowering weeks, and write down every flowering plant and every insect that lands. 358 surveys in all, from the first August to the last.
The totals came out lopsided in an interesting way.
Total insect abundance roughly tripled. The number of distinct insect groups rose by about 150 percent. Flowering plant species richness went up sevenfold.
And native bees, counted on their own, multiplied by a factor of 20.
That gap is the finding. A tripling of insects overall alongside a twentyfold rise in one group means the habitat was not simply getting busier, it was selecting for the animals that work flowers.
The part that happened outside the fence
None of that would matter much to a farmer if it stayed on the site.
So the team counted bees on the soybean flowers next door as well, and found twice as many visiting the rows closest to the solar fence as the rows farther away. Visitation beside the arrays came out on par with soybean grown beside conservation grassland, the land a federal program pays farmers to keep out of production.
Which is the quiet economic point. One neighbor is being paid to maintain that habitat, the other built it as ground cover under a power plant, and the crop next door cannot tell the difference.
It also lands differently than most arguments about panels on farmland, where the usual frame is land taken out of production rather than a service handed over the fence for free.
What this does not prove yet
It is two sites in one state, in a region where prairie grows anyway. Nobody has shown the same thing in the desert Southwest, and the researchers do not claim it.
Site preparation decides a great deal of it. An array built on a scraped and leveled pad starts from bare subsoil, and a seed mix put down there is beginning a much longer job than one sown into soil left in place.
The economics still push the other way. Pollinator seed costs more than turf and needs different mowing, and developers who might do it want to know they will break even.
There is a counterweight worth keeping in view too. Panels reflect polarized light in a way that can read as water to some insects, which is a real effect and not one this study measured.
A senior director at a clean energy policy center said that making solar farms support native pollinators is an obvious solution. On this evidence it is obvious, cheap and slow, and the slow part is the reason it keeps losing arguments to gravel.
Hugo is an engineer with strong technical expertise. Multilingual from an early age, his writing combines technical clarity with a strong interest in science and energy.
