The map of the Dolomites

A relief computed from real terrain elevation, with four hundred and fifty-nine places on true coordinates on top. Twenty groups: pick one to go inside it.

How it is built, and why you can trust it

The base is not a photograph. It is a physical map computed from a digital terrain model at roughly twenty-six metres per cell: a raking light from the north-west and three fill lights for the relief, valley incision, and elevation bands running from the green of the valley floors to the tan of the high pastures. Where OpenStreetMap marks bare rock the palette shifts to the pale grey of limestone, which is why the twenty Dolomite groups can be told apart at a glance. On top of it are the real rivers and lakes, again from OpenStreetMap, and the boundaries of the three provinces drawn from the actual administrative relations rather than traced by hand.

The map opens on northern Italy and after three seconds descends on its own onto the Dolomites: it is there to show where they are before you look at them closely. The cities on the wide view are only for orientation. Woodland is its own layer, from OpenStreetMap: without it the elevation bands alone would not say whether that green is pasture or forest, and it thins out with height as it really does. Lakes carry a pale rim so they read even when small — the Carezza is two hundred and fifty metres across, and at this scale it would vanish. The main roads are there too, from the valley motorways to the provincial roads climbing to the passes: on a mountain map they orient you better than anything else.

We built the satellite version and threw it away. From up there, under the forest canopy, almost nothing is distinguishable: the valleys become one flat dark mass and the shape of the mountain disappears. Shaded relief does the opposite, because it shows slope rather than colour.

The coordinates come from OpenStreetMap, queried across two boxes, the central Dolomites and the Brenta. The raw box included the Ötztal Alps, which are not Dolomites: the Wildspitze at 3,768 metres was the proof of the error. The filter that fixed it is geological and checkable: the Marmolada, with Punta Penia at 3,343 metres, is the highest peak in the Dolomites, so anything higher is out by definition.

Then the check against an independent source. Nine well-known places compared with Wikipedia’s coordinates: eight agree within eight hundred metres, most within twenty-five to four hundred. The one large gap is the Pordoi pass, and it is Wikipedia that is wrong, placing it 1.8 km east of the actual col.

The names of the twenty groups are not placed by eye: each is anchored to its own real peaks, and the label sits at their average. The method also discarded what did not hold. Sciliar, Schiara, Vette Feltrine and the Friulian Dolomites have no peaks in our dataset and so do not appear, rather than being placed from memory. And one case of shared naming nearly moved Fanes fifteen kilometres: the data contains a “Sasso delle Nove” near the Sella that is not the one in the Fanes group.

Villages link to their pages only where the name matches exactly: thirty-seven out of forty-one. Approximate matching was tried and dropped, because it produces plausible errors — it put the Pale di San Martino plateau on San Martino in Badia, forty kilometres away in another valley. A map with pins in the wrong place is worse than no map.

343 peaks and 20 groups, and no mark was drawn by hand: everything you see is data.