A bifacial panel is only half-installed when the modules are bolted down, because the whole point of paying for glass on both sides is the light that comes up from underneath — and that depends entirely on what you put on the ground. This is the part most backyard installs ignore, and in my case the fix was quite literally stacked on pallets along the back fence of my shop.
What bifacial gain actually is
A bifacial module makes power from its back face as well as its front. The front collects direct sun; the back collects whatever light bounces up off the ground below and behind the array. How much the back contributes is driven by one unglamorous number: albedo, the fraction of light the ground reflects.
The relationship is close to linear, which makes the decision easy to reason about. Bare soil or dry grass reflects about 0.2. Grey gravel isn’t much better. A white surface — clean light gravel or reflective fabric — runs 0.5 or higher. Fresh snow is up around 0.8. Push the ground under your array from 0.2 to roughly 0.5 and you can take the bifacial contribution from a token few percent up into the low-to-mid teens. On my system that pencils out to somewhere around +6–8% of total production, on the order of +1,000–1,300 kWh a year — every year, for the life of the array, for the cost of moving some rock around.
My panels help here too: they’re Canadian Solar CS3W-415PB-AG glass-glass modules with about a 70% bifaciality factor, meaning the back face is unusually productive and the array is more sensitive to ground albedo than a typical bifacial. High bifaciality plus a bright ground is where this trick earns its keep.
The ground that matters is behind and beneath, not out front
Here’s the geometry that trips people up. On a south-facing tilted array, the rear of the modules faces north and down. So the ground the back faces “see” is the strip directly under the rack and the apron immediately behind it (the north side). The sunny ground out in front? The back can’t see it. So you spend your reflective material under and behind the array, and skip the front entirely.
There’s a second twist that saved me a lot of hauling: my array sits low — bottom edge about two feet off the ground, average panel height around four and a half feet. A low array can’t “see” far, so the reflective apron doesn’t need to sprawl. For my geometry the sweet spot is the full footprint under the rack — roughly 13 feet deep on a 20° tilt — plus about 8 feet of apron behind it. Over a 54-foot row that’s around 1,100 square feet of ground to brighten. Going deeper than 8 feet would be wasted effort at this mounting height. (If your array is tall, you get to extend the apron farther and the whole effect is stronger — height is the lever most people underestimate.)
The part where my old shop paid off
My building used to be a marble and granite cutting operation, and I still have a lot of stone offcuts on pallets. Light marble is a genuinely good diffuse reflector — matte white marble sits right in that 0.5–0.6 albedo range. So instead of buying white gravel (there’s none local anyway; the Columbia Gorge is basalt country, and basalt is grey), I’m paving the reflective zone with offcuts I already own. If you’re not sitting on a pile of marble, the buy-it equivalents are crushed white limestone, dolomite, or white marble chips from a decorative-stone yard — the local grey aggregate won’t do much.
A few things I learned sorting through it:
Lay the matte, sawn face up — not the polished side. This matters more than it sounds. Matte stone reflects diffusely, scattering light in every direction so it feeds the back of the panels all day long. Polished marble reflects specularly, like a mirror — it throws a single bright beam that mostly misses the modules, causes glare, and turns into a sky-mirror when it’s wet. Rough side up wins.
Sort by color and give the dark stone a job. The whitest marble goes on the reflective surface. The darker granite offcuts — low albedo, useless as reflectors — become my perimeter anchors and wind ballast, where reflectivity doesn’t matter.
Lay it tight. Dark gaps between pieces are albedo leaking away, so crazy-paving it close pays off.
Fabric plus stone is the belt-and-suspenders version
Even with a yard full of marble, I’m putting a layer of reflective white/black woven ground cover (DeWitt orchard fabric, white face up) down first, as the base. It does two things stone alone can’t: it fills the gaps between irregular offcuts with a uniform white, and its black backing blocks weeds — and nothing tanks your local albedo faster than grass creeping up between the rocks. The marble then goes on top as the durable surface and the ballast that pins the fabric down. If you’re buying everything, fabric-plus-a-thinner-gravel-layer is often cheaper and lighter than deep stone, at a slightly higher albedo.
Orchardists here in Hood River already roll out this exact reflective fabric to color up apples and cherries by bouncing light into the lower canopy — same physics as my panel backs, which is a nice reminder that none of this is exotic.
Don’t Expect Miracles
Two caveats so nobody’s disappointed. First, those numbers are a planning estimate; real bifacial gain depends on your latitude, how cloudy your winters are, how much ground you actually cover, and how clean you keep the surface — light stone gets dusty and dusty stone reflects less, so it needs an occasional rinse — happens automatically when I wash the panels. Second, a low mount caps the ceiling no matter how white the ground is; if you have any say in the design, mounting the array higher does more for bifacial gain than any surface treatment. But as a retrofit to an array that’s already in the ground, a bright apron is close to free money, and mine came off a pallet.
And there’s a winter bonus: when it snows, the ground albedo jumps to 0.8-plus on its own, and the array gets a free bifacial spike on the cold clear days after a storm — as long as the snow’s on the ground and off the panel faces.
The reflective-apron cheat sheet
- Cover under + behind (north), never the front. The back faces see the ground beneath and behind the array.
- Size it to your mounting height. Low array (mine ~4.5 ft avg): full under-footprint (~13 ft on a 20° tilt) plus ~8 ft of apron — about 1,100 sq ft over a 54-ft row. Taller array: extend the apron and expect more gain.
- Aim for ~0.5 albedo. Matte white marble/limestone or reflective fabric ≈ 0.5; grey basalt ≈ 0.2; fabric alone 0.6–0.8; snow 0.8+.
- Matte/sawn face up, not polished — diffuse reflection beats a mirror beam and avoids glare.
- Sort light stone for the surface; dark granite for anchors. Lay it tight to kill dark gaps.
- Fabric base + stone ballast = uniform albedo plus weed control, using the stone you’ve got.
- Expected payoff (my 12.45 kW array, ~70% bifaciality, Hood River): roughly +6–8%, on the order of +1,000–1,300 kWh/yr. Planning estimate, not a guarantee.
- Maintenance: rinse the dust occasionally; keep weeds out from under the array.