Honeypot sites

Why does a national park funnel its visitors into a handful of crowded spots instead of spreading them thinly?

A park authority with a large tract of protected land and millions of visits a year responds by building four car parks, four cafes, four sets of toilets and four hard-surfaced paths, and putting every one of them on the map. The predictable result is that most of those visits land on a few hectares, worn to gravel and full of people, while the rest of the park sees almost nobody.

Say it as an accusation and it sounds damning: an organisation whose purpose is protecting land has deliberately concentrated the damage. The instinctive alternative is obvious — spread the visitors, and let each place carry a load so light it hardly notices. Why would anyone whose job is conservation choose the opposite?

The spreading intuition contains a hidden assumption, and the whole argument turns on it. It assumes damage is roughly proportional to use — that a hundred people in one place do a hundred times the harm of one, so a hundred places with one person each come to the same total. If that were so, concentration would buy nothing and cost a great deal.

So test the assumption. Walk once across a patch of upland grass and almost nothing happens. Walk a hundred times and the vegetation is bruised, stems broken, the sward thinning. Walk a thousand times and it is bare compacted soil. Now the question that decides the policy: what does the ten-thousandth passage do? Almost nothing, because there is nothing left to damage, and the marginal footstep costs the site essentially zero.

That is a curve that rises steeply and then flattens, and recreation ecologists have found it repeatedly: a large share of the vegetation loss at a site occurs at comparatively low levels of use, after which further use adds little. The finding is robust in shape though not in numbers — how fast a site saturates depends enormously on soil, slope, vegetation and season, and I would not quote a threshold.

Follow that shape through. Take some number of visits and two ways to place them. Concentrate them and you produce a few sites near the flat end of the curve — ruined, but only those, and their ruin nearly capped however many more arrive. Spread them and you place many sites on the steep part, each paying the expensive early damage. Where the curve is concave the totals are not close: spreading can do more aggregate harm with the same number of boots. Concentration is not damage traded for convenience; on this curve it is the option that produces less damage.

There is a second consequence of the same concentration. Hardening is affordable only where use is dense: a boardwalk, a stone-pitched path, a drainage scheme, a toilet block, a ranger — each has a fixed cost that is trivial per visitor at a hundred thousand visits a year and absurd at two hundred. So concentration does not merely cap the damage; it funds the engineering that caps it further, and a path built to take feet stops the feet taking the hillside. It also puts visitors within reach of a warning, a count, or a fee.

Now the honest correction, and it deflates the story of deliberate design. Most visitors concentrate anyway. The overwhelming determinant of where people go in a large protected area is where the road and the car park are, followed by how far they will walk — which for most is not far. A park authority that "chose" its honeypots largely inherited them from a road built for other reasons a century earlier. The management decision is usually the smaller one: given that the crowd will gather here regardless, do we harden this place and signpost it, or leave it to erode informally?

The argument rests entirely on the shape of one curve, so it is refuted by finding that shape wrong. If the impact-use relation were straight, concentration would be neutral; if it were convex — damage accelerating with use — spreading would be correct and the honeypot an error.

And there are real cases of exactly that. Put enough feet on a steep slope with erodible soil and you do not reach a bare stable surface; you expose a channel, water finds it, and a gully begins that deepens on its own long after the walkers have gone. That is accelerating damage, and it is why the same authority that funnels people onto a lakeshore will close and re-route a ridge path. Wildlife disturbance behaves differently again: the harm is per-encounter and does not saturate, so a bird that abandons a nest does not care that the site was already busy. Honeypotting is a policy for damage that saturates, and should be expected to fail where that does not hold.

Two further caveats. The honeypot is a genuine sacrifice: the sites chosen really are degraded, and calling them "sacrifice areas" is more honest than calling them managed. And concentration produces problems that are not ecological at all — traffic, parking overspill, the local village that has become a car park — which the ecological case does not answer.

It is also worth separating this from the crowding problems it resembles. When a much-loved town is overwhelmed by visitors, concentration is the disease and dispersal a proposed cure. Here the sign is reversed, and the reason is what is being protected: a town's resident community and visitor experience both degrade further the more people arrive, while a trampled hectare stops getting worse once the vegetation has gone.

• Why footpath erosion on slopes behaves so differently from trampling on flat ground, and how re-routing is decided.

Every term the collection defines is gathered in the glossary.