HomeMost solar in New York City goes on flat roofs, and every flat-roof project faces the same fork: hold the racking down with weight, or fasten it through the roof. Installers often present this as a preference. It rarely is. The structure and the membrane usually decide it, and understanding why tells you whether the answer you have been given is sound.
What each approach actually is
Ballasted racking sits on the roof surface and is held in place by concrete blocks. Nothing goes through the membrane. The array resists wind because it is heavy and because the racking is shaped to reduce uplift.
Attached (penetrating) racking is bolted through the roof into the structure below, with each penetration flashed and sealed. The array resists wind because it is mechanically fixed. It needs far less weight.
Both are standard, well-established, and code-compliant when engineered properly. Neither is a compromise.
The dead load question
Ballast is heavy. A ballasted array adds a few pounds per square foot across the roof, and more at the corners and edges where uplift forces are highest and more weight is needed to counter them.
On a modern concrete-deck building that is negligible. On the older housing stock in Brooklyn, Queens and the Bronx — wood joists, sometimes a century old, occasionally with previous alterations nobody recorded — it is a genuine structural question. Not automatically a problem, but genuinely a question.
Which is why a ballasted design on an older building should come with a structural assessment behind it. If an installer proposes ballast on a brownstone without having anyone look at the framing, they have skipped the step that matters. And a filing involving ballast without a structural basis is a filing that can come back at plan review, which costs you weeks.
When the structure cannot carry ballast, the design moves to attached racking. That is the normal resolution, not a fallback.
Wind, and why the edges are different
Wind uplift is the force actually trying to move your array, and it is not uniform across a roof.
Air flowing over a building separates at the edges and corners, creating suction that is substantially stronger there than in the middle of the roof. Engineering standards divide a roof into interior, edge and corner zones for exactly this reason, and the required ballast or attachment density rises sharply as you move outward.
Practical consequences:
- Perimeter areas cost more per panel — more ballast or more attachment points for the same output.
- Parapets help. A tall parapet disrupts the airflow that creates edge suction, reducing uplift on the array behind it. This is a real advantage of New York’s row-house roofs and one reason ballasted systems work well here.
- Height and exposure matter. Wind speed rises with elevation and with open exposure. A six-storey building near the water in Bay Ridge or the Rockaways sees materially higher loading than a two-storey house in Midwood.
- Panel tilt is a wind variable. A steeper tilt catches more wind, so it needs more restraint. This is part of why low-tilt layouts are common on city roofs — it is not only about landmark sightlines.
The design should be engineered for the loading at your specific site, not to a generic city figure. That calculation is part of the DOB filing anyway, so it exists — you can ask to see it.
The roof warranty, which is often the deciding factor
For many owners this settles the argument before engineering does.
If your roof is under warranty, penetrating it may void or limit that warranty depending on the terms and on who performs the work. Ballasted racking avoids the issue entirely by not breaching the membrane.
Where penetration is the right engineering answer on a warrantied roof, the workable route is usually to have the original roofer — or an installer approved under that warranty — perform and certify the flashing. That preserves coverage. It requires coordinating two trades, which is a scheduling matter, not a technical obstacle.
Get the roofer’s written position rather than the solar installer’s opinion about it. These are different companies with different liabilities, and only one of them is on the hook for the roof.
Membrane type matters more than people expect
New York flat roofs are typically EPDM rubber, TPO, or modified bitumen, and they behave differently under an array.
The universal issue with ballasted systems is abrasion. Blocks and racking feet resting on a membrane, with thermal cycling moving things microscopically through every day and night, will wear the surface over decades. Proper installation puts protective slip sheets or pads under every bearing point. This is basic and standard, and it is also the sort of detail a rushed crew skips because nothing goes wrong for eight years.
Ask specifically what goes between the ballast and your roof. It is a fair question and a revealing one.
Drainage, the failure mode nobody photographs
An array changes how water and debris move across a roof. Racking creates lines that catch leaves; ballast blocks create small dams. On a New York roof with limited drains and a plane tree overhead, this is how a functioning roof turns into a ponding roof.
Two things to insist on: the layout must not obstruct drains or scuppers, and there must be a way to physically reach them for clearing. An array you cannot walk to the drain through is a maintenance problem you have built in permanently.
This also connects to the fire-access clearances New York requires. Those clear paths exist for firefighters, but they double as maintenance access — one more reason the layout constraints are not pure loss.
How the decision usually resolves
In practice, on New York City flat roofs:
- Ballast tends to win on a sound roof with capacity to spare, an intact warranty you want preserved, and a decent parapet. It is the common answer on well-built low-rise buildings.
- Attachment tends to win on older or lighter framing that cannot take the dead load, on high or exposed buildings where uplift demands would require impractical amounts of ballast, and on roofs being replaced anyway — where penetrations can be integrated into new membrane work cleanly.
That last case is worth emphasising, because it points at the sequencing question underneath all of this. If your roof is near the end of its life, doing the roof and the solar as one coordinated job removes the warranty conflict, removes the abrasion worry, and avoids paying to remove and reinstall an array in five years.
If you want a straight answer for your roof — including whether the framing can take ballast at all — we will look at the structure before proposing a racking type, rather than the other way around. Any installer who names the racking before seeing the framing is guessing.