Viewshed: what an antenna really sees

The line of sight answers a precise question: does A see B? But whoever chooses the site of an antenna, a fire lookout or a wind turbine asks the reverse: from this point, what can be seen — and what can't? Answering line by line would take hours. The viewshed does it in one go, for every square metre around the point.

Québec 3D's new Explorer tool computes this viewshed on the LiDAR relief, accounting for forest, buildings and Earth curvature. It is free, needs no account, and everything is computed in your browser.

What is a viewshed?

It is the set of places visible from an observation point, within a given radius. The name echoes the watershed: where a watershed gathers everything that drains toward a point, a viewshed gathers everything that point “sees”. It is the basic tool of radio coverage studies, visual impact assessments and land surveillance.

Compute a viewshed in four steps

1. In the map panel, click “Viewshed”.
2. Set the height of the observation point above ground (30 m for a tower, for instance), the target height (1.5 m: a hand-held phone; 1.7 m: a walker's eyes) and the analysis radius, up to 20 km.
3. Click on the map at the chosen location (on a phone, tap it). The disc fills with colour in a second or two.
4. Click elsewhere to move the point: comparing several candidate sites takes a few clicks. Changing a height or the radius reruns the computation.

Three colours, three situations

• Green — visible: a target at the chosen height above ground is in direct view of the observation point.
• Amber — canopy reached: at ground level the target is hidden by what immediately surrounds it — the trees of a forest, the buildings of a street — but the top of that cover, the treetop or the roof, is in view. This is not a terrain shadow: it is a local screen. For a radio wave the distinction matters: the signal reaches the cover and passes through it, weakened, whereas a terrain shadow cuts it off.
• Shaded — hidden: a hill, a wooded ridge or a more distant obstacle stands in the way. Neither the target nor its cover is in view.

Amber only appears when “Surface (buildings + forest)” is ticked — the default. Unticked, the tool considers bare earth only: useful to isolate the effect of terrain alone, but too optimistic to judge real-world coverage.

Reading the results

The summary gives the share of the disc that is visible and the share where the canopy is reached, the visible area in km², the elevation of the observation point and the computation resolution. Below it, a curve shows the visible share by distance: hovering over the histogram reads the visible percentage ring by ring — and shows at a glance whether coverage collapses beyond a ridge, or picks up again on a facing slope. The CSV button exports these values, with the parameters and the data attribution.

How it is computed

The tool first assembles an elevation grid covering the disc: bare earth and surface (ground plus canopy and building height). It then casts rays from the observation point, one per pixel of circumference — several thousand. Along each ray it keeps the highest elevation angle of the obstacles already crossed: that is the horizon. A target is visible if it sits above this horizon; the top of the cover, if it rises above it, marks the amber zone. Earth curvature and atmospheric refraction are included with the standard coefficient k = 4/3, as for the line of sight.

The resolution adapts to the radius to keep the computation fast: about 1.6 m per cell for a 1 km radius, 6.5 m for 5 km, 13 m for 10 km and 26 m for 20 km (in southern Québec). For a detailed look at one area, a smaller radius is therefore better.

Why so little green in towns and forests?

The first surprise is often the same: from a 30 m mast in the middle of a neighbourhood or a forest, the disc is almost entirely shaded. This is not an error. Trees 15 to 25 m tall, or buildings, close the horizon within a few hundred metres; a target 1.5 m above the ground, at the foot of other trees or between other buildings, then almost always sits below that horizon. In the middle of the Plains of Abraham in Québec City, a 30 m mast thus sees about 1% of a 3 km disc at phone height — versus 63% if forest and buildings are ignored.

This is exactly why cellular networks do not rely on direct view alone: waves bend around obstacles by diffraction, pass through foliage and walls while weakening, and bounce off façades. The viewshed does not model these phenomena; it shows the geometry they act upon. The amber zone shows where the signal arrives above the obstacles, and the shade, where it will have to rely on diffraction.

What it is good for

Cellular telecommunications — Compare candidate tower sites, measure what an extra 10 or 20 metres of height buys, spot the shadow zones of an existing site, plan densification by targeting the sectors the terrain hides. The viewshed is a quick first screening, before radio prediction tools.

Radio and private networks — VHF/UHF relays for emergency services, LoRa gateways, municipal or agricultural networks: from where does a relay cover the most ground?

Wind energy and landscape — Thanks to the reciprocity of sight, the viewshed computed from a wind turbine's hub, with a 1.7 m target, shows from where the turbine will be seen. The same reasoning applies to a tower, a silo or a tall building.

Surveillance and safety — Siting a forest-fire lookout, a camera or an observation post: the viewshed reveals the blind spots.

Tourism and the outdoors — What can you see from this lookout, this summit, this observation tower? With an observer height of 1.7 m, the tool draws the panorama on the map.

Limits worth knowing

This is geometric visibility, not a coverage prediction. The tool computes neither diffraction, nor loss through vegetation and buildings, nor reflections, nor transmitter power. The surface is smoothed and, at large radii, the resolution coarsens: a narrow obstacle (isolated tree, tower, power line) can slip through. The forest cover dates from the LiDAR flight: later cutting and regrowth are not reflected. Heights are measured from the ground, and real refraction varies around k = 4/3 with the weather. To size a network, these results call for a proper radio study.

The takeaway

The viewshed turns a field question — what can we see from here? — into a map. On Québec's LiDAR relief, forest and buildings included, it separates what is in view, what is hidden only by its own cover, and what the terrain truly masks. Together with the elevation profile and the line of sight, it completes a free toolkit for reading the land in three dimensions.