Building Downwards: The Case For Underground Cities

AThe argument for putting part of a city underground is almost always made in the language of land. Space at street level in a successful city is scarce and dear, the reasoning goes, so whatever does not need daylight ought to be moved beneath it and the surface released for something better. This strikes me as the least persuasive of the arguments available, and it is the one that gets made. The price of land does rise steeply, but so does the cost of removing what lies under it, and the second curve is the steeper of the two: one basement level is inexpensive, a third is not, and a fourth may cost more than the site above it. Land prices are in any case the most volatile quantity in urban economics, whereas a tunnel lasts a century. Justifying a permanent structure by a temporary price is a habit that ought to have died in 1974, when three ambitious schemes in one country were abandoned half finished for precisely that reason.

BPetra Obradovic, who has spent two decades costing excavation for the Sorvik Institute of Subsurface Planning, has put numbers on the second curve. In sound rock, she reports, a cubic metre of usable space can be won for about 190 euros; in the saturated clays that underlie most river cities the same cubic metre costs between 700 and 900 euros once ground freezing, continuous pumping and the monitoring of buildings overhead have been paid for. The difference is not a detail of engineering but the whole economics of the question, since the cities under the greatest pressure to expand are exactly those that sit on estuaries and floodplains. Obradovic maintains that the geological survey, not the land registry, ought to decide whether a scheme is even considered. Her rule of thumb is blunt: beyond 30 metres, or in anything softer than firm marl, the answer above ground will nearly always be cheaper.

CThe case that does survive scrutiny is a duller one. Temperature ten metres down barely moves across the year, so anything that has to be held at a constant temperature, whether an archive, a cold store or the halls of servers that no city admits to needing, belongs there on thermal grounds alone, and the saving is not trivial: the municipal archive in Sorvik reports 62 per cent less demand for cooling than the surface building it replaced in 2016. Freight is the second candidate. Where deliveries are gathered at a terminal on the edge of a city and carried in beneath the streets, lorry movements in the centre drop sharply, and the scheme in Duvenholm reduced them by 27 per cent within two years of opening. Neither use obliges anybody to spend a working life below ground, and that, in my view, is the distinction that ought to organise the whole debate.

DThe distinction is regularly blurred by the objection raised most often, which is that people dislike being underground. Surveys support the claim in a loose way: asked in 2019 whether they would accept an office two floors beneath the pavement, 68 per cent of respondents across four cities said they would not. Rolf Teichman, an environmental psychologist, dismisses such surveys as worthless; what respondents reject, he argues, is a windowless room rather than a buried one, and the two have been confounded because architects built them together for a century. In his own trials, daylight was carried down light wells into offices below the pavement, and satisfaction scores came out indistinguishable from those recorded on the third floor, provided that the wells were kept clear of obstruction. The trials ran for six weeks, however, and six weeks is not a career. Whether the effect would last remains, on the published evidence, unknown.

EA further obstacle lies beyond the reach of engineering altogether. In many jurisdictions whoever owns the surface owns everything beneath it to an unstated depth, so a tunnel has to be negotiated plot by plot with every proprietor whose land it passes under, and one refusal can move a route half a kilometre sideways. The countries that have built most successfully below their cities are, without exception, those in which statute ends private ownership at a fixed depth: the national code in Sorvik fixed that depth at 25 metres in 2003, and the effect on approvals was immediate, the average time falling from 51 months to 19. Reform of this kind is unglamorous, costs nothing to enact and is opposed by nobody who can be named in advance. That so few countries have attempted it is, I think, the clearest sign that the subject is not being taken seriously.

FMy own position is accordingly narrower than the enthusiasts allow and wider than the sceptics will admit. Build downwards for things rather than for people: for the archive, the reservoir, the pipe and the pallet, none of which cares about daylight and all of which are grateful for a steady temperature, and do it where the ground is sound and the law is clear. Build downwards for people only where the alternative is plainly worse, as it is for a crossing beneath a river or a platform beneath a square. The general case for underground cities, conducted in the language of land prices, is a weak case that happens to point in a useful direction; the particular cases, conducted in the language of temperature, freight and statute, are strong ones that nobody ever puts on a poster. It is unfortunate that the weak case is the one that raises money.