How the two differ in kind
These two arrive looking like a question of values and turn into a question of statics within about ten minutes. It is worth being precise about what each one physically is. A green roof is a permanent layered assembly covering the plan area of the roof: a root-resistant membrane or root barrier, a drainage and reservoir layer, a filter fleece, the growing medium, and the planting. Its design weight is not what it weighs on the day it goes down. It is what it weighs when the medium is holding all the water it can hold, because that is the condition it will reach, repeatedly, for as long as the building stands. A ballasted solar array is a different animal entirely — a racking frame that is not attached to the building at all. It stays on the roof because it is heavy enough not to leave, which means its weight is not a property of the product but the deliberate output of a wind calculation. Nobody wanted that mass up there; it is there to win an argument.
That difference propagates into everything the engineer has to check. The green roof's load is uniformly distributed, present all the time, and additive with whatever snow sits on top of it — the question is whether the deck, the joists and everything below them can carry another dead load spread across the whole covered area. The array's load is concentrated at rack feet and ballast blocks, and it is deliberately unequal across the roof, because uplift is not uniform: corners are worse than perimeters and perimeters are worse than the field, so the ballast plan is a map rather than a number. The array also introduces a load case the green roof does not have at all — a net upward one. And it brings a downward case of its own that is easy to forget: tilted rows are obstructions, obstructions catch drifting snow, and the drift that collects against a row of modules is a design case in its own right.
Then comes the asymmetry that usually decides it. If the structure cannot take the ballast, the array has somewhere to go: fix the racking to the structure mechanically or with bonded mounts, and the counterweight largely disappears, trading dead load for roof penetrations, a detailing job and a conversation with whoever warrants the membrane. Hybrid layouts do this selectively, attaching only the corner and perimeter zones where the ballast demand is steepest. If the structure cannot take the growing medium, a green roof has far less room to move. Lightweight engineered media and a shallow extensive profile buy something real, but the saturated case is not optional and there is a depth below which the planting cannot survive a dry summer, at which point what you have is a maintenance liability with a green roof's paperwork. So the honest way to run this decision is against four questions, in this order: how much spare capacity the structure has and whether the limit is a global one or a local one; whether the membrane has just been laid or has years left on it; how much of the roof survives shading, row spacing and the setbacks the fire service expects; and whether this building needs a return it can meter or can bank a benefit it cannot.
The factors that actually differ
Show
| Vegetated green roof | Ballasted solar array | |
|---|---|---|
| The load case that governs | Downward, and only downward. A permanent distributed dead load at its fully saturated value — the design condition, not the installed one — with snow sitting on top of it. There is a wind problem, but it is about the medium scouring at edges and corners before the plants knit, and it is solved with aggregate margins, erosion control and getting establishment right, not by adding weight. | Both directions at once. Wind uplift is the reason the ballast exists, and the ballast is then a downward load the structure never asked for. Tilted rows add a second downward case by catching drifting snow. Remove the wind argument and the entire weight problem is negotiable. |
| How the load reaches the structure | Spread. Depth multiplied by saturated density gives a pressure that applies across the whole covered plan area, so every member under it shares in the same proportion. The check is a global one about the roof as a whole. | Concentrated and uneven by design. Load arrives at rack feet and ballast blocks, and the layout puts more of it in corners and along perimeters than in the field. The check is local and repeated: these members, under these points, plus whether a paver is sitting where a joist is not. |
| Whether the load is negotiable | Barely. A lighter medium and a shallower profile are the only levers, and both run out — a depth that will not carry a plant through drought is not a green roof. The saturated weight is a permanent condition, not a worst case you can manage away with drainage. | Very. Mechanical or bonded attachment converts most of the ballast into fixings, and a hybrid layout attaches only the zones where the demand is worst. The price is penetrations, detailing and a membrane warranty conversation — but there is a route through a structural no, and the other option does not have one. |
| What happens to the waterproofing | It is buried, permanently, under a rooted assembly. That demands a root-resistant specification and a leak test before anything covers it, because afterwards a leak is located by excavation rather than by walking the roof. On larger installations a permanent leak-detection layer under the build-up is the standard answer to exactly this problem. | It is covered, not buried. Protection mats go under the ballast and the rails, abrasion and point pressure are the failure modes to detail against, and the roof stays walkable and inspectable between the rows. A leak is still found the ordinary way. |
| When the roof needs replacing | The green roof is the demolition phase of that reroof. The medium comes off wet, at its heaviest, in the least convenient possible form, and it does not simply go back down afterwards. In practice this chains the green roof's life to the membrane's, which is why it is a new-roof decision and a poor bolt-on to a roof with a decade left. | A planned, expensive interruption. De-ballast, lift or slide the racks, reroof, put it back — a sequence racking manufacturers design for, and one that costs generation time as well as labour. Unpleasant and budgetable, which is a materially different thing from irreversible. |
| How much of the roof you actually get to use | Nearly all of it. What you lose is vegetation-free margins at parapets, around penetrations and to outlets, and access routes for maintenance. Shade does not remove area — it changes the species list. A roof overshadowed by a taller neighbour is still a perfectly good green roof. | Much less than the plan area suggests. Row spacing to stop the rows shading each other, setbacks and access pathways, clearance from plant and stair cores, and the shadow of anything taller than the array. On a cut-up roof the difference between roof area and array area is the whole business case. |
| Small and edge-dominated roofs | Penalised gently. Cost tracks area but the detailing tracks perimeter, so a small or awkwardly shaped roof pays proportionally more for edging, margins and outlet details, and gets proportionally less growing area for it. | Penalised hard, and in the direction that hurts. When almost the whole roof is inside the perimeter and corner wind zones, the ballast demanded per module climbs steeply — a small roof is the case where ballasting fails on load and mechanical attachment stops being optional. |
| What it asks of you every year | Horticulture, permanently. Irrigation through establishment, weeding, replanting failures, and keeping outlets and inspection chambers clear. It is the only one of the two where neglect creates a new problem rather than just lost value: a drought-stressed, weed-colonised roof is a fuel load and a drainage risk. | Electrical and monitored. The array reports its own underperformance, which is a genuine advantage — you know when something is wrong without going up there. Cleaning is occasional and climate-dependent, and the inverter is a known mid-life replacement rather than a surprise. |
| What it gives back | A benefits case, mostly unmetered: stormwater retained and peak runoff delayed, a membrane shielded from UV and from daily thermal cycling, summer heat gain buffered, biodiversity, and amenity if anyone can see or use it. Where a local authority credits retention or requires it, that unmetered benefit acquires a value — which is worth checking before the decision, not after. | A metered output. It generates, the generation is measurable, and it can therefore be modelled, financed and argued about on a spreadsheet. Whatever the local export and incentive regime is, it applies to a number somebody can read off a meter. |
Which one, and when
Choose vegetated green roof when…
- The membrane is new, or is being replaced right now. This is the only moment at which burying waterproofing under a permanent assembly is a sound decision, and it does not come round again for decades.
- The roof is overlooked or occupied — flats above it, a taller neighbour, a terrace people use. Amenity and outlook are worth something here that an array cannot deliver, and an ugly roof is a visible one.
- Runoff is the problem you are actually solving, or retention and biodiversity are conditions of consent rather than aspirations in a design statement.
- Shading, orientation or a roof cut up by plant and stair cores would reduce an array to a token gesture. Shade changes a green roof's planting list; it destroys an array's yield.
- The spare capacity is generous and genuinely global — the structure can take a distributed dead load everywhere, which is exactly the shape of load this option applies.
Choose ballasted solar array when…
- The spare capacity is thin, unknown, or belongs to a structure nobody wants to open up. You need the option that survives a structural no, and only one of these has a way out of the load.
- The roof is large, simple and sunlit, with few obstructions to cast shadows or eat into row spacing — the geometry an array converts into output most efficiently.
- The case has to be made on a metered return that can be modelled and financed, rather than on a benefit the organisation would have to agree to value.
- The membrane has years left but not decades. An array can be lifted and put back at reroof; a green roof cannot, and installing one now spends the rest of that membrane's life on your behalf.
- Nobody in the building's operation is going to do horticulture. There is no irrigation, no maintenance regime, and no one whose job it will be to weed a roof in five years' time.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Can I do both — solar panels over a green roof?
- Yes, and biosolar roofs are an established combination rather than a compromise. The pairing is genuinely complementary in two ways: the vegetation keeps the roof surface cooler than a membrane or gravel would, and modules lose efficiency as they get hotter, so the planting does the array a small favour; and in systems built for it, the green roof build-up itself provides part or all of the array's ballast, which means the two loads overlap instead of simply adding. That last point is the one to test early, because the alternative is that they do simply add — and a roof that could carry either on its own may well refuse both. The other conditions are practical and are usually what kills a badly planned biosolar roof: rows have to be spaced and raised enough that a person can get underneath to weed and inspect, the planting under the modules and along their shaded side is in permanent shade and needs a species list that reflects that, and the maintenance regime has to be written for a roof where the horticultural work and the electrical work interfere with each other. Ask about it at concept stage, when the structural grid and the array layout can still be drawn around each other.
- Which one actually weighs more?
- There is no general answer, and anyone offering one is not looking at your roof. The green roof's load falls out of two things you choose — how deep the medium is and what that specific medium weighs when fully saturated, which is a product property from a data sheet and varies widely between lightweight engineered media and anything approaching native soil. The array's ballast falls out of things you largely do not choose: the design uplift pressure at your location and building height, and which roof zone each module sits in. The one dial you do turn is the safety factor, and turning it down is a judgement about how much uncertainty in the wind assumptions and in ballast friction you are willing to carry, not a way of making the roof lighter. That is why a small, exposed roof can demand more ballast per module than a large sheltered one for an identical array. The more useful point is that the total is the wrong comparison anyway. One load is spread evenly across the whole area and the other lands at points and clusters in corners, so two options with the same total mass can produce completely different answers from the same engineer. Run both calculators to get the two quantities in front of you, then hand over the distribution as well as the number.
- Do these calculators tell me whether my roof can take it?
- No, and both say so in their own notes. The green roof calculator multiplies depth by area by saturated density to give a total weight, and converts it to the pressure figure an engineer works in — it is a load quantity, and it knows nothing about your structure. The ballast calculator turns a design uplift pressure and an array area into the counterweight that resists it with a safety factor applied, using a simplified whole-array tributary approach; it states plainly that real uplift varies significantly by roof zone and that the final ballast layout normally comes from the racking manufacturer's engineer for your specific roof. What both are for is arriving at the structural conversation with the right things in hand: the quantity you are proposing to add, expressed as a pressure, and where on the roof it lands. What the engineer then needs from you is the rest of it — the existing roof build-up's own dead load, the snow and live loads the structure was designed for, what the original design allowed as spare, and whether anything has been added since the drawings were made.
- Does a green roof really make the membrane last longer?
- The mechanism is real and well understood: the two things that age a single-ply membrane hardest are ultraviolet light and the daily thermal cycling of a black surface in the sun, and a green roof removes the first almost entirely and flattens the second substantially. The membrane underneath lives in a mild, stable, shaded environment. The awkward part is that you cannot see it, and the whole benefit is contingent on the installation being right the day it was covered — which is why the sequence matters so much more than on an ordinary roof. Flood test or electronically test the membrane before a single layer goes over it, use a root-resistant specification rather than assuming the standard one will hold, and on anything sizeable install a permanent leak-detection layer so that a future leak is a location on a plan rather than a search. Get that sequence right and the argument holds. Get it wrong and you have converted a repairable roof into an excavation, which is the single most expensive way this decision can go wrong.
- Which one costs more?
- They are expensive in different shapes, which is why no figure belongs on this page. A green roof is material-and-area heavy: barrier, drainage layer, filter, medium and planting all scale with the area covered, and there is a substantial one-off logistics cost in getting many tonnes of medium up onto a roof by crane or blower that has nothing to do with the materials themselves. Its detailing scales with perimeter, so awkward shapes cost more than their area suggests. Then it carries a maintenance cost that never ends and cannot be deferred indefinitely without losing the asset. An array is equipment-and-design heavy: modules, racking and inverters, plus electrical labour, a wind analysis and a ballast layout, with a known inverter replacement in the middle of its life. The structural difference that matters more than any of that is the direction of the two cash flows — one of these produces a metered output and the other produces benefits your organisation has to agree to value. And the cost of changing your mind is wildly asymmetric: taking an array off is a bad week, while taking a green roof off is a demolition of wet material at its heaviest. Price both against your own roof and your own quotes, and price the reversal too.
- What do the two do to roof drainage?
- Opposite things, and both need designing rather than assuming. A green roof deliberately intervenes in the hydrology: it retains rainfall in the medium and the reservoir layer and releases the rest slowly, which is one of its strongest arguments and, in some jurisdictions, the one that pays. The obligation it creates is permanent access — outlets need inspection chambers that stay reachable through the build-up forever, and a blocked outlet under a planted roof is a problem discovered late. An array changes runoff volume not at all, but it can absolutely change where the water goes: ballast blocks and rails laid without reference to the roof's falls will dam flow across a surface designed to move it, and a ballast layout that puts weight over or beside an outlet creates ponding on a roof that drained perfectly well the week before. On both, the rule is the same and is easy to state and easy to forget — the drainage design existed before your new roof use did, and it is your job to leave it working.