Roads and landslide terrain.

What updates and what does not. The map, the kilometres and the colours below are static: they come from official maps that rarely change, and we redo them only when the Region revises them. Only the "Recent rainfall" box is live and refreshes on every visit to the page.

Recent rainfall

We compare the rain of the last hours (Open-Meteo model, at the "Atri, Calanchi" station point) with the scientific landslide-triggering curves for Abruzzo (Brunetti et al., 2010, 5% threshold: E = 5.94 · D0.45, E in mm and D in hours). Exceeding it does not mean anything will slide: it means that, in the past, rainfall of this size has been associated with landslides in the region.

Model or rain gauge? The rain above is calculated by a weather model, not measured on the spot. A real rain gauge exists: the station of the Calanchi Reserve (MeteoTeramo network) publishes its data. We do not show it on this page because it belongs to another body and we have not asked permission to reuse it; every day, however, we compare it with the model to check how far off it is. Official data from the regional network are held by the Abruzzo Functional Centre (AllarMeteo).

Map

The colour of each 100-metre stretch shows how much the 50-metre strip around the road overlaps terrain already mapped as landslide-affected, steep or eroded. It does not indicate ongoing movement, nor that the road is unstable. Many mapped areas are old or dormant.

Roads with the most kilometres in mapped areas

What is not in this table. Unnamed dirt tracks (country tracks) are the most exposed group by far: 26.2 km out of 170.6 cross mapped areas, more than all named roads put together (16.6 km out of 174.6). We do not rank them because they have no name, and they are not coloured on the map (the index excludes them); but they count in the totals above. Source of the counts: OpenStreetMap crossed with the official landslide maps.

Is the ground moving? Radar satellites

The European service EGMS (Copernicus) measures from radar satellites by how many millimetres a year thousands of fixed points, such as roofs, walls and asphalt, rise or sink. We looked at the official viewer (2020-2024 edition): in built-up areas there are many points; where the ground is bare or covered in vegetation, as in the badlands, the radar finds no reliable points and few or none remain. "No point" does not mean "no movement": it means the satellite does not measure there.

For this reason we do not yet report a value for individual roads: a few hand readings would not say anything about a stretch, and we prefer not to make things up. The road-by-road calculation needs the official data to be downloaded (it requires a European EU Login account) and is the next step. EGMS data are updated once a year: they cannot tell what is happening today.

How it is computed

• Landslide areas: IFFI inventory (ISPRA) and the Abruzzo Region geomorphological map, "landslide movements" layer.
• Slope: Copernicus GLO-30 terrain model (30 m). It underestimates scarps and roads on embankments.
• Badlands and erosion: regional geomorphological map and modelled erosion (Abruzzo Region).
• Roads: OpenStreetMap, excluding farm tracks.
• Index: 0.30·slope + 0.35·landslide area + 0.15·badlands + 0.20·erosion, weights chosen by judgement and not calibrated on events. Low below 0.25; medium from 0.25; high from 0.5.

Limits

The maps are official but not updated in real time; the rain is from a model, not a rain gauge; the index is not calibrated and is not a probability. In past data, the two certain dates of heavy rain with road landslides (15 November 2017 and 1-3 April 2026) exceeded our thresholds: a good sign, but two cases prove nothing.

Seen a road collapse or landslide?

Report it by marking the spot on the map and describing what you saw (photos cannot be attached yet): it helps check whether the map matches reality. Go to Reports

Data: roads · mapped areas. Processed on 2026-10-20. In an emergency call 112.