- Where the water is even a candidate
- What is actually gained, and what is only plausible
- The electrical plant is mostly the land plant, on a longer cable
- When to refuse the site
- Frequently asked questions
- Do floating panels always produce more than land panels?
- Can the inverter float with the modules?
- Does floating solar save a known amount of water?
- Sources and references
Floating solar is a photovoltaic array on a buoyant structure over water, not a land array that happens to be near a pond. The modules are ordinary enough that a National Renewable Energy Laboratory assessment describes them as adapted from land systems. The structure, the anchors, and the cable to shore are not. Inverters, that assessment says, typically stay on shore, with power brought across by floating or underwater cable. A design that puts the inverter on the water because a brochure showed it there has left the arrangement the assessment described, and it owes an explanation the brochure did not provide.
The benefits people quote are real in some places and unmeasured in others. The same assessment says studies have documented power gains from cooler modules over a wide range, about 1.5 to 22 percent, and it says comprehensive cost data for the claim of cheaper land and site prep are lacking. A gain printed without the study, the climate, and the mounting is not a yield you can finance. Evaporation savings and effects on algae are discussed in the literature the assessment reviews, with an explicit uncertainty: less light may reduce algae and may also affect other aquatic life. “Good for the reservoir” is not a finding.
Where the water is even a candidate
NREL’s later summary of its own siting rules is a filter, not a wish list. The area needs to be free of swift current, free of ice flows and heavy snow load, and free of freight traffic whose wakes would work the moorings. If the water body sometimes dries, the bottom has to be flat enough for the floats to ground. The site should pose minimal additional environmental risk. A 2025 National Laboratory of the Rockies note adds the plain version: some reservoirs have ship wakes that can damage lines and floats, and some are too cold, too shallow, or too steep on the bottom to hold the array.
The US Bureau of Reclamation’s 2021 considerations report is stricter where a dam is involved. Energy yield has to include cooling, which can help, and changing water levels, which can tilt the array or set it on an uneven bottom and cut output. Evaporation benefits, the report says, still need study, and a small system on a large reservoir may not change evaporation in a way you can count. Dam-safety review is mandatory in that agency’s process if the installation could affect the dam. A floating array is not approved because the modules are listed.
What is actually gained, and what is only plausible
| Claim | What the opened sources support | What they do not support |
|---|---|---|
| Uses water instead of new land | The NLR note and the NREL assessment both treat this as the point, on water bodies that pass the filters above. | A right to cover a lake that has traffic, ice, or a dam-safety conflict. |
| Cooler modules, higher output | The assessment cites documented gains of about 1.5 to 22 percent across studies. | A single percentage for every reservoir. |
| Less evaporation | Reclamation says the effect needs further study and may be negligible for a small cover on a large lake. | A water-rights credit you can book without a measurement. |
| Cheaper than land | The assessment says anecdotal claims exist and comprehensive cost data do not. | A conclusion that floating is the low-cost option. |
Salt water and some process ponds add corrosion. The assessment says those sites need different materials. Anchors and moorings have to tolerate level changes, wind, and waves. Systems are sometimes designed to rest on the bottom if the reservoir is drawn down. That is a structural requirement, not a feature you discover during the first dry year.
The electrical plant is mostly the land plant, on a longer cable
Because the assessment places inverters on shore, the grid rules do not become optional. A grid-tied inverter still has to cease energizing an island within 2 seconds under IEEE 1547-2018, as the NREL primer states for US interconnection. The unusual parts are the floating or underwater cable, the mooring loads on that cable, and access for maintenance that may need a boat. Soiling still happens. Pollen, birds, and dust do not stop at the shoreline. PNNL’s advice that soiling is often invisible, and that wash frequency depends on rain, tilt, and dust, still applies. Doing the wash from a rocking platform is an access cost the land array did not have.
A fishery that wants panels over working water is a biological project as well as an electrical one. Existing notes on ponds and aeration are in fishery and solar hybrids. Do not treat a generation assessment as a stocking study.
When to refuse the site
Refuse it, or redesign it, if current, ice, snow, or ship wakes are part of the year. Refuse it if drawdown will ground the array on a slope the floats cannot tolerate. Refuse it if a dam owner cannot complete the safety review Reclamation describes. Refuse a yield model that uses 22 percent extra energy without showing which study and which climate produced that end of the range. Prefer a land array when the water fails those tests. Floating solar is a way to use a qualifying surface. It is not a way to skip one.
Frequently asked questions
Do floating panels always produce more than land panels?
No. The NREL assessment reports documented cooling gains of about 1.5 to 22 percent across studies, which is a range, not a guarantee. Water level changes can also cut output if the array tilts or grounds. Model the site.
Can the inverter float with the modules?
The assessment says inverters typically stay on shore, with cable out to the array. A floating inverter is a different design and needs its own justification, especially for access and corrosion.
Does floating solar save a known amount of water?
Reclamation says evaporation effects still need study, and a small system on a large reservoir may not matter. Do not enter an unmeasured volume into a water budget.
Sources and references
Device arrangement, the 1.5 to 22 percent study range, evaporation and ecology caveats, and missing cost data: NREL floating photovoltaic technical potential, OSTI 1489330. Siting filters: NREL international FPV applications. Wakes, cold, depth, and slope: National Laboratory of the Rockies, 2025. Dam safety, yield, and evaporation: US Bureau of Reclamation, 2021. Soiling on any array: Pacific Northwest National Laboratory. Shore inverters and islanding: NREL/Sandia IEEE 1547-2018 primer.
