Coverage in detail

The coverage map shows where aircraft were observed without radio signal. This page looks at one square of it, up to 40 km on a side, and asks why. For every point of the square it checks which ground stations of the Open Glider Network have a straight line to an aircraft flying there at a chosen height above the ground, and then estimates the share of the time the aircraft goes unheard, from how often a station at that distance was measured to receive that kind of aircraft. Where no station has a line of sight, the cause is the terrain, and only a new antenna in a better place would help. The same estimate works where nothing has been measured yet, because nobody there flies with two radio systems or no station is near, and it is the place to start when a club considers a receiver. Back to the coverage map.

The square

Pick the kind of aircraft, its height above the ground and the size of the square, then click the map to centre it somewhere else. The 3D view shows the same result draped on the terrain, and a click on it draws the lines of sight from that point to the stations that see it.

What the monitor observed here

The dashed outlines on the map are the 0.25-degree cells of the coverage map that overlap the square. Each is labelled with the share of the time that the probe aircraft, those whose losses come mostly from coverage, flew without signal in the band of height that holds the chosen height. The table sets that share beside the mean of the model's estimate over the part of the cell that lies in the square. Where the monitor measured a cell, trust the measurement.

What the colours assume

The ground stations are those of the Open Glider Network that were heard in the last week and have forwarded aircraft to it, the 40 nearest within 100 km of the centre at most. The altitude a station reports in its beacon is often wrong (the stations page flags dozens), so each antenna is taken as 10 m above the terrain at the station. The terrain comes from Mapterhorn’s open height data, loaded by your browser: about 100 to 200 m cells for the surroundings and about 20 to 40 m cells over the square. The earth is a sphere of 4/3 of its radius, the usual allowance for the bending of radio waves in the air, and the ground is the only obstacle: the model leaves out trees, buildings and masts. The number of points of the square is chosen to keep the work bounded.

Each station in line of sight of a point receives an aircraft there with a probability that depends on the distance and on what the aircraft carries. It was measured station by station and minute by minute on 6 to 29 October 2026, counting only the OGN stations the model uses, for stations that could see the aircraft. For a paraglider or hang glider with FLARM it is 93% within 2 km, 77% at 5 to 10 km, 46% at 15 to 20 km, 25% at 30 to 40 km and 3% at 75 to 100 km. The pilot's body shields a paraglider's radio in every turn, and that is in the measurement. With FANET the figures are lower near the station (74% within 2 km) and close beyond 30 km. Gliders and powered aircraft fall from about 95% within 2 km to 28-31% at 20 to 30 km. FANET ground stations near flying sites were left out, because many of them receive poorly: FANET within 2 km was heard 29% of the time by FANET ground stations and 75% by OGN stations. Between the bins the curve is interpolated on the logarithm of the distance, and beyond 100 km it is zero. The probability that no station receives the aircraft is the product of one minus each probability, and that is the estimate of the time without signal. A point that no station sees is at 100%.

Where a measurement exists it is shown, in a solid tint, and the model fills the rest in a paler tint with a diagonal hatch. A cell with no value, a phone cell nobody measured, is left untinted, and in the measured-only view the edge of the measured area is outlined. Colours: red is an estimate over 60% of the time without signal, amber 40 to 60%, light amber 20 to 40%, light green under 20%. The estimate was compared with the coverage measured in 340 cells: it is pessimistic by about 15 points on average, and it ranks the cells only partly right (rank correlation 0.45, where the measurements agree with themselves at 0.86). Where the monitor measured a cell (the dashed outlines), trust the measurement. On Monte Grappa the line of sight alone put two or more stations in view over most of the square. The probe flying there was without signal 70%, 69% and 61% of the time at 0 to 300, 300 to 600 and 600 to 1,200 m above the ground, and this estimate gives 75%, 62% and 52%.

The measured volume comes from the same October 2026 feed: the aircraft positions of 6 to 29 October 2026, the time without signal spread along each gap in an aircraft's track, and the seconds airborne, all counted per voxel. A voxel is about 2 km by 2 km by 100 m high (0.02 degrees) or, as a fallback, 5 km by 5 km by 250 m (0.05 degrees), in altitude above the sea, from 0 to 6,000 m. It is shown only if at least three aircraft-days and ten minutes of flying lie in it, and the finer voxel is used when it passes that floor, else the coarser. The model is used only where no voxel passes: measured wins where it exists. The channel is either radio (FLARM, ADS-L, OGN trackers, PilotAware and FANET, also when an internet gateway forwards them) or phone apps (SafeSky, Naviter, Vololibero, Microtrak and flyxc, which reach the network over the internet). There is no model for the phone channel yet, so where no voxel passes the map is left untinted. The families are paragliders and hang gliders on the one side and gliders and powered aircraft on the other; drones are left out. In the 3D view, “Volume” draws these voxels as translucent boxes at their altitude, solid and edged, coloured by the share without signal, with a select to slice it: the chosen height above the ground plus or minus 150 m, all altitudes, or a band of 500 m. For the radio channel it also draws the model’s estimate as faint boxes without edge, very faint green where the estimate is under 20% without signal (covered, drawn on the outer surface only; the select can leave them out to show the shadows alone), on a grid of 48 by 48 columns and levels of 250 m up to 4,000 m or 2,000 m above the highest ground, wherever no measured voxel covers the centre of the box; a click on a box tells which it is, its value and, for an estimated one, the stations in sight. The estimated volume uses the same model as the colours. Checked against the measured volume up to 4,000 m (coarse voxels with at least three aircraft-days and ten minutes, Alps, 6 to 29 October 2026): for gliders and powered aircraft the estimate ranks the voxels about as well at 2,000 m above the ground as near it, and is about 10 points optimistic at every height; above 2,000 m it gets the level right within about 10 points but not which voxels are worse. For paragliders, measured up to 2,000 m above the ground, the estimate ranks as well aloft as low down but becomes more optimistic with height, about 16 points at 1,200 to 2,000 m. Where a voxel is measured, the measurement is shown. The phone channel has no estimated volume yet. For now the tiles cover the Alps and their surroundings (43 to 48.5 N, 5 to 17 E); elsewhere only the model is shown. Credit as for the rest of the page: it contains data from the Open Glider Network.

Limits. The measurements are from October, when fewer aircraft fly and the stations may differ from summer. The body shielding is included in the measured reception and is not modelled separately. Aircraft antennas and station antennas are averaged, so a single aircraft with a poor installation, or a station with a better mast, departs from the curve. For powered aircraft the reach was measured and the estimate was not validated against measured cells, because the method has probe aircraft only for free flight and for gliders. Helicopters were not measured.