What is a rooftop farm good for? — between 633 hectares and 1,284

This series has handed out, one at a time, the tools for seeing a roof as farmland. The first instalment counted the area, the second took the physics of weight and water down to the level of design, the third traced the genealogy, the sixth followed the allocation of cost. The tools are assembled. But the better assembled they are, the more conspicuous the question we kept deferring: what, then, is a rooftop farm good for? Both defenders and critics tend to swap that question for another. The defenders say it is good for a city to have more greenery; the critics say you cannot feed a city that way. Neither is wrong, and neither is answering the other. They fail to meet because nobody settled, before the argument began, what the roof was supposed to deliver.

This article splits the expectation into three and tests each against figures. First, food: how much does a rooftop farm actually put on a city's table? Second, environment: is growing it overhead really lighter than trucking it in? Third, green space policy: has the duty to green a roof become an alibi for the greenery lost at ground level? On all three the answer moves towards not as much as advertised. Then we retrieve a fourth use — one nobody was expecting — from a school roof in Rome. Finally, granting every one of the limits, we set out in order the first steps a citizen can take. Counting limits is not the same as giving up. Only once you have settled what a thing cannot do can you settle what it can.

One thing should be said in advance. Of the seven days planned, this series could not publish its fourth instalment, on cases around the world, or its fifth, on where Japan stands. This article therefore does not sum up seven pieces of argument; it stands on the five that were actually published — area, physics, genealogy, cost, and this one. What those two missing instalments would have covered, above all the present state of Japan's roofs, is filled in here only as far as figures we could verify allow, and that is no substitute. Having chosen not to rewrite the shape of the series after the fact, we record the gap as a gap.

Since 2000 Japan's Ministry of Land, Infrastructure, Transport and Tourism has run an annual survey of rooftop and wall greening work nationwide. Its latest results, published on 12 December 2025, put newly constructed green roofs in 2024 at about 14.5 hectares and green walls at about 4.6. Cumulatively, from 2000 to 2024, that is roughly 633 hectares of roof and 131 of wall. The first instalment of this series cited the then-current figure of about 615 hectares through 2023; adding another year does not change the character of the number. Twenty-four years have accumulated 633 hectares, and the most recent year added 14.5 — the growth, as the first instalment showed, has already levelled off.

Now set those 633 hectares beside the ground. According to the agriculture ministry's account of Japan's allotment gardens, fiscal 2024 counted 4,273 gardens nationwide, 188,713 plots and 1,284 hectares. By operator, 1,977 were opened by local authorities, 1,433 by farmers, 449 by companies and non-profits, and 414 by agricultural cooperatives. Divide 1,284 hectares by 188,713 plots and each plot comes to about 68 square metres — that much soil in the hands of each of more than 180,000 plot-holders. In fiscal 1992 the same figures stood at 691 gardens, 56,727 plots and 202 hectares, so the area has multiplied roughly six and a half times in a little over thirty years.

Side by side, two things emerge. The gardens on the ground hold about twice the area of twenty-four years of rooftop greening; and, decisively, all 1,284 hectares of it is soil meant to be eaten from. Of the 633 hectares overhead, most is thin-layer greening of sedum and the like, and how much carries edible crops these statistics do not say. The comparison has limits, of course. Allotment gardens are a nationwide figure that includes flat suburban land, while a green roof can exist in dense central districts where nothing else can — a property no area figure captures. Even so, if the object is simply to multiply the places in a city where a person can touch soil, these two numbers have very nearly answered which route is efficient.

Arguments about how much a rooftop farm contributes to food usually begin from an assumption: if every suitable roof were used. Assumptions make for circular debate, so let us take instead a single roof that has actually been measured over a long period. On the roof of the ICTA-UAB building at the Autonomous University of Barcelona sits a 125-square-metre greenhouse integrated into the structure, where yields and resource use were recorded for every tomato cycle from 2015 to 2023. The report by Evangelista and colleagues, published in Agronomy for Sustainable Development in 2026, gathers those eight years together. A research building's roof, run by trained technicians, measured continuously while the growing conditions were varied — as records of rooftop farming go, this is close to best case.

Under those best-case conditions, the yield came out as a range: 0.9 to 20.3 kilograms per square metre. The best cycles reached 49 grams per plant per day; others fell to less than a twentieth of that. And across the eight years the trend was a decline of 31.2 percent. The cause was neither disease nor climate but the degradation of the polycarbonate sheeting over the greenhouse. Transmissivity falls, light falls, yield falls. The authors recommend replacing the sheets every four to six years rather than the manufacturer's ten, reporting a 19.4 to 31.8 percent gain in productivity and a 6.7 to 7.7 percent reduction in global warming impact when it is done. A rooftop farm is not a facility that can be described by its performance on the day it was built. Farming on a roof declines at the same rate as the roof itself ages.

That variance is fatal in the context of food supply, because a city plans its food not on averages but on guaranteed minimums. A facility capable of a 0.9-kilogram year cannot be counted as part of supply. And as the sixth instalment of this series showed, neither Brooklyn Grange nor Lufa Farms makes its profit from the harvest itself: one sells the technique of building roofs, the other sells distribution. The operators themselves treat what the roof produces as something other than food supply. That we should stop appraising rooftop farming as food production is the most candid conclusion these eight years of measurement will support.

The second reason usually given for rooftop farming is transport. Shorten the distance from field to table and you burn less fuel and lighten the load on the environment — an intuitive claim, and one rarely tested. The same eight-year Barcelona record ran a life cycle assessment alongside the harvests. It found the global warming impact per kilogram of tomato ranging from 0.54 kilograms of CO2 equivalent in the lightest cycle to 12.05 in the heaviest. Same roof, same greenhouse, same crop. Where does a spread of more than twentyfold come from? Artificial lighting. Cycles using lamps consumed 19.7 kilowatt-hours per kilogram; ordinary greenhouse cycles, 0.49. A factor of forty.

What that figure means is that what governs a rooftop farm's environmental performance is not where it sits but how it is run. Being on a roof in the city centre contributes almost nothing to the calculation. What contributes is whether you burn lamps, whether you heat, when you replace the covering, whether you recirculate water and nutrients. The authors note that recirculation reduces losses of both but demands careful monitoring to hold nutrient concentrations where they should be. A rooftop farm's environmental performance, in other words, resides in the quality of daily management rather than in the design. That sits exactly at the junction of the physics discussed in our second instalment and the money discussed in our sixth: the moment the money for upkeep runs out, the environmental performance falls with it.

Reservations are needed. These figures come from a research-grade integrated rooftop greenhouse in Barcelona's Mediterranean climate, and the authors themselves write that cases at higher latitudes may require additional inputs and shift the assessment. They also state explicitly that one of the artificially lit cycles, with eighteen plants, was not representative and cannot be extrapolated. So one cannot say that the carbon intensity of rooftop farming runs from 0.54 to 12.05. What one can say is that within a single facility it can move by more than an order of magnitude with operating conditions, and that environmental performance cannot be described by the single metric of transport distance. The method set out in our first instalment — never to conclude from one indicator — holds here too.

Of the criticisms aimed at rooftop greening, the heaviest is the charge of an alibi: build out to the edge of the site, erase the greenery at ground level, and make up for it overhead. The regulatory greening ratio is satisfied while the greenery visible from the street, and the trees rooted in the earth, diminish. How far does that charge hold as a matter of regulation? Tokyo's rules let us check. Article 6 of the enforcement regulation under the Ordinance on the Protection and Restoration of Nature in Tokyo sets the greening standard in two tracks. For ground level, Table 2 gives site area minus building footprint, multiplied by 0.2 to 0.25. In addition, Table 3 sets a frontage requirement: the length of the site's road frontage multiplied by three tenths to eight tenths. For building tops, Table 4 gives the roof area multiplied by 0.2 to 0.35.

What matters is that the two are set separately. Article 14, paragraph 1 of the ordinance itself speaks of a plan for greening both at ground level and on the building, and the structure is not one in which greening a roof lowers the ground-level requirement. In principle, then, the alibi does not hold. But Article 6 of the regulation continues. Paragraph 3 provides that where there is a special reason making it difficult to meet the ground-level area standard, the difficult portion may be replaced by planting trees over an equal area on the building. Paragraph 4 provides the reverse, allowing a building-top shortfall to be made up at ground level. A route for substitution exists in the rules, and it is open in both directions.

The honest answer to the criticism, then, is that the rules do not ordinarily permit the alibi, but the door to the exception stands open — and how wide it stands cannot be read off the text. What counts as a special reason making compliance difficult, and how many filings have actually used the substitution, could not be verified from the environment in which this was written. Whether the criticism lands is a question for operational statistics rather than for the text, and those statistics we do not yet have. The pattern from the sixth instalment — that the intent of a rule and the direction the money runs do not coincide — recurs here as the relation between a rule's text and its administration. It is worth noting that the separate frontage standard, imposed on top of the ground-level one, shows Tokyo's system trying to protect the greenery visible from the street as a thing in its own right; on the alibi charge, that is the regulation's own defence.

So far we have pared three expectations back with figures. Let us set down, last, a use nobody expected. On the roof of the Keplero scientific high school in Rome sits a green roof of about 200 square metres installed in 2016 — 150 square metres of thin-layer system on a 12-centimetre substrate, and 50 square metres of experimental raised beds at 15 centimetres. Six Mediterranean species were planted at installation. As Bellini and colleagues reported in the journal Plants in 2025, this roof, left largely untended thereafter, had been colonised by 62 plant species from 18 families. Thirty-five of them, 56 percent of the total, were annuals.

Six species became sixty-two, and nobody planted the difference. The authors observe that native Mediterranean species can serve as a valid alternative to sedum carpets, and that creating varied microclimatic conditions supports natural colonisation and can lead to more resilient and more biodiverse roof systems. What the result suggests is that a roof's surest product may not be the harvest at all. As the sixth instalment showed, of a green roof's benefits only rainwater and the cooling bill carry a price; habitat for living things carries no label. Yet what this roof actually produced was the unlabelled kind — and producing it required not investment but being left alone.

The authors state their own reservations plainly: the vegetation survey ran for one year, and plants that seem at first to thrive can later disappear. Sixty-two species cannot therefore be read as this roof's permanent condition. Nor can a result from Rome's Mediterranean climate be carried unaltered onto a Japanese roof with its rainy season and its typhoons. Still, the question the case raises remains: might it sometimes be worth more to a city to lay a substrate, shape some microtopography and leave the rest to whatever arrives, than to spend labour and money making a roof grow crops? That question, at the last, unsettles the very premise on which this series has called the roof farmland.

The second instalment framed weight as the roof's constraint, the sixth framed money. Seen from a citizen's side, however, the constraint that actually binds is simpler: the roof is locked. Without the permission of the building's owner or manager, the last step of the stairs cannot be climbed. In this the institutional character differs entirely from an allotment garden on the ground. For allotments, four statutory routes govern who may open one — the Citizen Farm Development Promotion Act, the Specified Farmland Lending Act, the Urban Farmland Lending Act, and the farm-use arrangement. Four categories of operator are contemplated: municipalities, agricultural cooperatives, owners of farmland, and companies or non-profits that own none. A citizen who wants to take part has an institutionally settled place to go.

The gap shows even in the existence of statistics. For allotment gardens, the agriculture ministry has tracked the number of gardens, plots and hectares consistently from fiscal 1992 to fiscal 2024, publishing the breakdown by operator. For rooftop farms there is no comparable national count. What the transport ministry's survey counts is the area of greening work, without distinguishing whether crops are grown on it or whether citizens can get to it. What is not counted is hard to make policy about. The sixth instalment said that a benefit with no price is not counted when the cutting starts; here the problem sits one step earlier. A place that is not counted never reaches the table where the argument happens.

The asymmetry does not refute the possibility of rooftop farming. It does reorder the priorities. If the aim is more chances for citizens to touch soil, opening more unlocked places is faster than negotiating open the locked ones. Ask the converse — what can only be done on a locked roof — and the answer is the one this series has given since day one: in dense central districts where the ground offers not one square metre to spare, a roof can still process rain and heat and serve as a corridor for living things. The value of a roof lies not where it competes with the ground but where the ground does not exist.

The first move is to start on the ground, not the roof. Many people who want to touch soil begin by looking for a roof, but Japan's 4,273 allotment gardens as of fiscal 2024 offer more than 188,000 plots averaging about 68 square metres. By operator: 1,977 opened by local authorities, 1,433 by farmers, 449 by companies and non-profits, 414 by agricultural cooperatives. Applications go through a municipality's agriculture desk or an agricultural cooperative, and this is the option to check before hunting for a roof. Building growing experience first bears directly on whether the roof works when you get there. As the second instalment showed, a roof is windy, its soil is shallow and its water drains fast — it is not a beginner's environment.

The second move, if you do choose a roof, is to verify the building before anything else. The order starts not from what you want to plant but from what the structure permits: year of completion, structural type, and whether a certificate of inspection exists under the Building Standards Act. The Mitaka grant scheme seen in the sixth instalment made it a condition that the building be strong enough to carry the new load and hold such a certificate, precisely because no institution will perform that check on your behalf. The third move is to find the water's exit first. What fails first on a roof is not the plants but the drainage. Locate the outlets, and confirm there is a route by which a person can reach them for inspection, before any soil goes up. When the second instalment said that a rooftop farm's drawings are read from the bottom, this was the order it meant.

The fourth move is to build the exit into the design. The cumulative area of rooftop greening is announced as about 633 hectares, but that figure is each year's construction summed, and how much of it has since been removed the survey does not say. Whether the accumulated area is the same as the area now actually in place, we do not know. On paper, in other words, rooftop greenery never disappears; in fact it does. When the soil goes up, write down under what conditions it comes down and who bears that cost. Whether the owner is a residents' association or a company, only the roofs where this was settled at the outset survive a change of the person in charge. Here is the reverse face of what the sixth instalment noticed about operators who buy no land: a place you borrow is a place you eventually give back.

This article tested, one by one, three expectations placed on rooftop farming, then set down a fourth use nobody expected, and finally laid out an order of practice. The route ran like this. First we set two areas side by side: about 633 hectares of green roof accumulated between 2000 and 2024, with roughly 14.5 hectares newly built in 2024, against 4,273 allotment gardens in fiscal 2024 comprising 188,713 plots over 1,284 hectares, an average of about 68 square metres each. Then we tested the contribution to food against eight years of measurement in a 125-square-metre rooftop greenhouse at the Autonomous University of Barcelona; then the environmental load against the life cycle assessment of the same facility; then the charge that rooftop greening serves as an alibi in green space policy, against the text of Tokyo's regulation. From there we moved to a school roof in Rome, and closed with four practical moves ordered from the ground upward.

Three points deserve restating by name. First, a rooftop farm's yield is unstable even under the best conditions. Across eight years in Barcelona it ranged from 0.9 to 20.3 kilograms per square metre and declined 31.2 percent overall, the cause being degradation of the covering material; the authors report that replacing it every four to six years improves productivity by 19.4 to 31.8 percent. Second, environmental performance is decided by operation, not location. Within the same facility, global warming impact ran from 0.54 to 12.05 kilograms of CO2 equivalent per kilogram, and electricity from 19.7 kilowatt-hours per kilogram under artificial light to 0.49 in ordinary cycles — a fortyfold gap. Third, Tokyo sets its rooftop and ground-level standards separately. Article 6 of the enforcement regulation gives ground level as site area minus building footprint times 0.2 to 0.25, and building tops as roof area times 0.2 to 0.35, with substitution permitted under paragraphs 3 and 4 only where a special reason makes compliance difficult, and only area for area.

The unknowns and reservations should be stated. The largest is that this series could not publish its fourth instalment, on cases around the world, or its fifth, on where Japan stands. What this article could show of Japan's rooftops is therefore a single kind of figure — the transport ministry's record of construction — and it has not addressed removals or actual operation. Whether the cumulative 633 hectares equals the area now in place, that survey does not say. The Barcelona figures come from a research facility in a Mediterranean climate; the authors state that higher latitudes may need additional inputs, and that one artificially lit cycle of eighteen plants was too small to extrapolate from. Rome's sixty-two species are the result of a one-year vegetation survey, and the authors caution that plants which seem at first to thrive can later disappear. How Tokyo administers its special reason making compliance difficult could not be verified. An objection is fair, too: that unstable yields are beside the point, since a garden's value in education, welfare and community should be measured on another scale entirely — and to that objection this article has offered no figures.

As the closing instalment of the series The roof as farmland, let us gather what seven days brought into view. On the first day we began by recounting the roof as area. The second took us down into the physics of weight and water, the third traced a genealogy that began with fire safety, the sixth brought the three instruments — mandate, subsidy and price — down to their actual sums. And this instalment has run all of it through the single question of what it is good for. What is left, having run it through, is that calling a roof farmland does not quite work. As a site of food production it is small and unstable; on environmental load it can be made worse by how it is run; and it is no substitute for greenery at ground level. Yet a roof can put a surface that processes rain and heat where no ground exists, and make an island where living things stop over. Sometimes, when the tending stops, sixty-two species arrive. What this article recommends a reader do next is to grow for one year in an allotment garden on the ground, and then go and check the inspection certificate and the drain outlets of their own building. The real uses of a roof come into view once we stop calling it farmland.