Green roof: blanket or habitat?

Does a green roof really serve biodiversity? Not always, to our mind: a sedum mat, however green it photographs, is not an ecosystem. It takes little substrate to hold it in place, a single family of succulents to make it flower in June, and no relief at all to give it life. A roof that is truly alive reads like a soil: depths that vary, hollows that hold water, refuges where something can hide. It is not the same roof, nor the same project.

The sedum mat has a real merit: it holds back a little rainwater, insulates a little in summer, greens an image. But laid as a thin, continuous layer, it houses no one. A ground beetle looking for the shade of a stone finds none; a solitary bee that digs into warm sand has no sand to dig into; a bird watching for a puddle after the rain finds only flat, dripping foliage, nothing to land on. Green monoculture is a roof that looks like a garden and behaves like a carpet.

What changes everything is relief. A roof designed for the living varies its substrate depth, thin in places, deep in others, and that variation alone is enough to grow different species zone by zone, dry ones on the mounds, cooler ones in the hollows. It carries materials that do nothing for waterproofing and everything for shelter: a pile of dead wood, a bed of sand, loose stacked stones, a puddle a few centimetres deep that dries out in summer and fills again in winter, like a real pond. Nothing spectacular. A diversity of small conditions, side by side, rather than one thin mat everywhere.

The law, here, already asks for more than a blanket. It requires, on a large share of new non-residential buildings, a green or solar roof: 40% of the surface as of this summer, 50% from next summer. And the text does not simply call for green: it asks for a system that favours, in so many words, “the preservation and reconquest of biodiversity”. A regulation that writes that phrase is not asking for a decorative mat laid down to tick a box; read closely, it is asking for what relief alone can give. The exact thresholds and scope must be checked project by project, never taken on faith, but the letter of the law already leans toward the living, not merely the green.

This picks up directly where we left the question of open ground: rebuilding a soil on a roof is never the first choice, only the one that remains once natural ground is out of reach. But since a structure has to carry it, waterproof it, sometimes water it, and have someone named to look after it, that rebuilt soil may as well be worth something. The gap in cost between a flat mat and a roof with relief lies mostly in the early design, not the site work: a landscape consultant brought in at sketch stage costs less than a waterproofing lot redone after the fact.

So the next time a roof plan lands on the table, one simple question before drawing: will this roof have relief, or will stay flat under its green? A mat unrolls. A habitat, by contrast, is drawn.