Two design choices add production to a model, and both depend on inputs that are easier to assume than to measure — which makes them the two places where an optimistic estimate most easily hides.
Teams reviewing yield assumptions can keep the engineering model separate from the record of review work; further details is one option for documenting time and handoffs.
Reviewed August 9, 2026.
Bifacial gain
Bifacial modules collect light on the rear face as well as the front. The additional yield is real and it depends entirely on how much light reaches the back.
For independent background on photovoltaic performance, resource data and current industry practice, consult PV Education.
Which depends on albedo — the reflectance of the ground under and around the array.
Albedo varies enormously. Fresh snow, light gravel, concrete, dry grass, dark soil and a black membrane roof are not close to each other, and the range between them changes the gain from substantial to negligible.
So the question is: what albedo was assumed, and was it measured? A regional default applied to a dark roof produces a gain in the model that will not appear on site.
Also relevant: mounting height, row spacing and how much of the rear face is shadowed by the structure itself. A bifacial module mounted tight to a dark roof is a monofacial module that cost more.
Trackers
Single-axis tracking follows the sun through the day, raising annual yield materially in the right conditions.
The gain is real and it is not free:
Moving parts. Maintenance and availability assumptions change, and a stuck tracker is a loss that monitoring may not flag.
Terrain. On uneven ground, tracker rows need either civil works or terrain-following capability, and the cost lands in one place or the other.
Self-shading. Trackers at low sun angles shade each other more than fixed arrays, and the modelling of that is a real variable.
And climate. The gain is largest in high direct-irradiance climates and smaller where the resource is diffuse — which means a figure quoted from a sunnier region does not transfer.
What to ask
For bifacial: "What albedo value, and where did it come from?" Measured, regional default, or manufacturer's marketing figure — three different answers.
"What is the modelled bifacial gain as a percentage?" It should be stated separately, not folded into the yield.
For trackers: "What availability is assumed, and does it account for mechanical downtime?" A 99% assumption carried over from a fixed array has ignored the moving parts.
And for both: "What does the model produce without this?" The difference is the value of the choice, and seeing it separately is how you judge whether the extra cost is worth it.
The general pattern
Both technologies add yield through an input that is hard to verify from a proposal.
That does not make them wrong choices — they are frequently right ones, and the gains are well established where the conditions support them.
It makes them the items to check most carefully, because they are where a plausible-looking model can differ most from the plant that gets built. The same logic as unsurveyed shading, applied to a decision that adds capital cost as well as modelled yield.
The short version
- Bifacial gain depends on albedo, which varies enormously between snow, gravel, concrete, grass, soil and a dark membrane roof
- A bifacial module mounted tight to a dark roof is a monofacial module that cost more
- Tracker gain is real and brings moving parts, terrain cost, increased self-shading, and a climate dependence that does not transfer between regions
- Ask what albedo was used and where it came from, and what availability a tracker model assumes for mechanical downtime
- Ask for the modelled gain as a separate percentage, and for the model output without the feature
- Both add yield through inputs that are hard to verify from a proposal, which makes them the items to check most carefully