Why Is the Sky Yellow, and What Is It Doing to Your Solar Panels?

If you looked up over New England on Tuesday afternoon and thought something was off, you weren’t wrong. The sky went from summer blue to a milky white to a sort of brownish-yellow over the course of a few hours, and the sun turned into a dull orange disc you could look at without squinting.

That’s Canadian wildfire smoke. And if you own solar panels, your monitoring app almost certainly noticed before you did.

Here’s what’s actually going on, what it means for your production numbers, and what’s worth doing about it. Spoiler on that last part: less than you’d think, but not nothing.

Where is this smoke even coming from?

Where is this smoke even coming from?

Not from anywhere close by, which is the part that surprises people.

There are currently more than 800 wildfires burning across Canada, including a growing cluster just north of the U.S. border near Minnesota. Upper-level winds lined up almost perfectly to carry the plume more than 1,000 miles, first across southern Canada and then hooking right, straight into New England. According to the Canadian Interagency Forest Fire Centre, 32 wildfires sparked across mostly western Ontario on Monday, with 46 burning out of control across the province.

The setup is a heat dome. Air flows clockwise around these big, slow-moving high-pressure systems, and this week’s is parked with its northern edge sitting right over northern Minnesota and southern Ontario where the fires are. That placement pushes the smoke east and south, into the Midwest and Northeast.

So nothing is burning in Massachusetts. The fires are a thousand-plus miles away and the smoke is doing the commuting.

What is the smoke actually doing to sunlight?

This is where it gets relevant to your roof.

Solar panels care about irradiance, which is the amount of solar energy landing on a given area of surface. There are two flavors of it worth knowing. Direct normal irradiance (DNI) is the light coming straight at you from the sun in a beam. Global horizontal irradiance (GHI) is everything that reaches a flat surface, the direct beam plus all the light that’s been scattered around the sky.

Put more simply: DNI is the sunbeam, GHI is the sunbeam plus the glow.

Smoke wrecks the first one a lot more than the second. A 2025 study in Nature Communications found that during days of thick smoke, clear-sky DNI in California declined 32–42%, while GHI only declined 11–17%. That gap matters, because GHI is the main resource your rooftop panels are running on.

Think of it like a lampshade. Put a thin white shade over a bare bulb and the room doesn’t go dark, the light just stops coming from one hard point and starts coming from everywhere at once. Your panels don’t need the beam specifically. They’ll take the glow. They just take less of it.

WBUR described the layer as acting like a giant, semi-translucent sponge, filtering and scattering the incoming solar radiation. That’s the mechanism, and it’s also why the sky looks the way it does. Smoke particles scatter the shorter wavelengths — the blues and violets — while the longer wavelengths, the reds and oranges, punch through. Same physics as a sunset, running at 2 p.m.

How much production are you actually losing?

Depends entirely on how far you are from the fire. That’s really the whole answer.

For systems near an active fire, it’s brutal. The Department of Energy cites research showing that during the September 2020 California wildfires, PV output declined 10% to 30% statewide but dropped as much as 58% locally. Across a set of monitored sites in the western U.S., wildfire smoke caused a measured drop of 9.4% to 37.8% in power production.

But we’re not near the fire. We’re the far end of a 1,000-mile plume, and that’s a meaningfully different situation. The Nature Communications team found that large GHI reductions are possible close to fires, but average GHI declines minimally — under 5% — from transported smoke. Their headline takeaway was that average PV resource losses stay modest outside the areas immediately near active fires, where plumes are thick and fresh.

One study modeled it city by city and the numbers are worth sitting with: during the heavy smoke year of 2020, panels near the fires in Modesto, California could have produced 11% less than normal, while farther out the effect shrank — Denver at 3%, New York at 1%.

That’s the annual figure though, and this is where people get confused. A 1% hit spread across a year is not the same as what your app shows you on a single bad afternoon. On the day, under a thick plume, the dip is real and you’ll see it. Averaged over twelve months, it mostly disappears into the noise.

So both things are true. Today stings. The year probably won’t.

One more wrinkle worth knowing, because it explains why you can’t just eyeball the air and guess: research from South Dakota Mines found that widespread smoke can cut individual panel output substantially even on days when the smoke is high up and ground-level air quality isn’t significantly affected. The smoke doesn’t have to be in your lungs to be in the way of your panels. Which is exactly what’s happening this week — in New England, most of the smoke is expected to stay at higher altitudes.

What does this do to the grid?

This is the underrated part of the story, and New England has been through it before.

When smoke blanketed the region in June 2023, ISO New England — the regional grid operator — saw it immediately. “At a simple level, the smoke is inhibiting the solar power from hitting the panels and generating electricity,” ISO-NE spokesperson Matt Kakley told WBUR at the time.

Here’s the twist. Because most solar in New England is “behind the meter” — on rooftops, used directly in the home — when production drops, demand on the regional grid goes up. But the smoke also held temperatures lower than forecast, so people ran their AC less and used less electricity than expected. Two conflicting things at once.

And ISO-NE’s forecasting models weren’t built for it. “When we forecast demand on the system, we’re looking at historical data,” Kakley said. “And the challenge with something like this is that there’s just simply not the historical data that you would have to look back on.”

ISO-NE’s own summer 2023 recap put it plainly: smoke from Canada significantly diminished the region’s solar production, the haze pushed temperatures below what models forecast, and together those two factors made forecasting grid demand quite challenging. Forecasters adjusted their modeling and the grid operated normally.

The scale has changed since then, which is why this keeps mattering more. New England had less than 80 MW of solar in 2010. By the end of 2025 the six-state region was at roughly 9,600 MW, producing an estimated 11,600 GWh that year. The region now has more than 8,000 MW of behind-the-meter PV, projected to cut peak-hour demand by more than 1,700 MW this summer.

You can see the same collision happening right now. The smoke absorbed enough of Tuesday’s peak heating to hold temperatures 5 to 10 degrees below forecast — Boston hit 90 against a forecast 97. CBS Boston lowered Wednesday’s forecast highs by several degrees and lifted its heat alert because of it. Less sun on the panels, less heat driving the AC. The grid feels both.

Okay, so what can you actually do?

Straight answer first: there is no trick that makes a panel see through smoke. Anyone selling you one is selling you something. But there are a few real moves, and a couple of things that are worth not doing.

Don’t clean your panels for this. This is the big one, and it’s where advice written for California gets misapplied out here. Cleaning helps when there’s physical ash sitting on the glass. Ash from wildfire smoke landing on panels is a real problem and hosing it off brings the system back toward peak production. But that’s for people who live near the fire and are getting actual fallout. We’re getting smoke aloft from a thousand miles away, not ashfall. There’s nothing on your glass to remove. Climbing on a hot roof in bad air to wash off a haze that’s in the atmosphere is all risk and no return.

Don’t panic at your monitoring app. A dip on a smoke day is your system working correctly. It’s reporting less sun because there’s less sun. One useful benchmark from a California installer: if you notice production has dropped more than 50%, that may be a technical issue or system failure rather than the weather. Short of that, a bad afternoon under a plume is expected behavior, not a fault.

Shift your heavy loads. If the middle of the day is producing less than normal, that’s the window where you’re pulling more from the grid instead of your roof. Dishwasher, laundry, EV charging — moving those to a clearer day or overnight is small, but it’s free and it’s the only lever most people have on a smoke day.

Watch the pattern, not the day. This is the one that actually matters, and it’s the reason to pay attention at all. CNN notes that with several months left in wildfire season, the door stays open for more Canadian smoke plumes to migrate south. The Globe pointed out that Tuesday’s haze was reminiscent of the last two summers, when heavy Canadian smoke drifted into the Northeast. This plume will move on. The pattern that produced it won’t.

Is this the case for a battery?

It’s at least worth thinking about, and here’s the honest version of why.

A battery doesn’t help your panels produce in smoke. Nothing does. What it does is decouple when you make power from when you use it. Your system banks the good hours and you draw on them during the bad ones instead of buying from the grid at the exact moment everyone else in the region is doing the same thing.

That’s not a smoke solution specifically. It’s an intermittency solution, and smoke is just this week’s version of intermittency. In 2023, ISO-NE noted the diminished solar production didn’t threaten grid reliability, partly because solar wasn’t yet a huge share of New England’s electricity. That’s less true every year. The more of the region’s power that comes off rooftops, the more a few smoky days in July actually move.

Storage is also the answer to a problem smoke days don’t cause but grid stress does: grid-connected solar panels without batteries don’t work if the power goes out.

The short version

• The smoke is from Ontario and Minnesota fires, carried 1,000+ miles by upper-level winds around a heat dome.
• It scatters sunlight rather than blocking it outright, so your panels lose some output but don’t stop.
• Near a fire, losses run 10–30% and can hit 58% locally. This far downwind, the annual average impact is a few percent or less.
• A single smoke afternoon still shows a real dip in your app. That’s normal, not a malfunction.
• Don’t clean your panels. We’re getting haze aloft, not ash on glass.
• Fire season runs for months. The plume leaves, the pattern stays.

Your system is fine. It’s doing exactly what a solar array does when there’s less sun: making less power, and then making full power again when the sky clears. That’s the whole story on the equipment side.

The longer story is that New England is now on its third straight summer of this, the region’s solar fleet has grown from 80 MW to nearly 10,000 MW in fifteen years, and the grid operator is openly saying its models don’t have the historical data for it. That’s not a reason to worry about your panels. It is a reason to think about where your power comes from at 3 p.m. on a yellow-sky day.

If you’re curious what that looks like for your specific roof — what you’re producing now, and whether storage makes sense for your usage — we’re happy to walk through the numbers with you. No pressure either way. Sometimes the answer is that you’re already in good shape.

---

_{Sources: WBUR, Boston Globe, CBS Boston, CNN, FOX Weather, Nature Communications, U.S. Department of Energy, ISO New England, Solar Power World, Palmetto.}