Terrain profile and path clearance
Point at two ends of a link and see what the ground in between does to it: the diffraction loss, the Fresnel zone clearance, and how much antenna height the path needs.
Endpoints
Link parameters
The map is drawn from the elevation tiles themselves, so it needs no external map service. A = site, B = far end; the dashed line is the path.
Terrain profile
How this is calculated
Elevation comes from AWS Terrain Tiles, Terrarium encoded: the height in metres is (R x 256 + G + B/256) - 32768. The data is SRTM at about 30 m over land, merged with GMTED and ETOPO1. For every sampled point the line of sight between the two antennas is compared with the ground after adding the earth bulge b = d1 x d2 / (12.75 x k), and the first Fresnel zone radius r1 = 17.32 x sqrt(d1 x d2 / (fGHz x (d1+d2))). The obstruction above the line of sight gives the knife-edge parameter nu = h x sqrt(2(d1+d2)/(lambda x d1 x d2)) and the loss L = 6.9 + 20 log10(sqrt((nu-0.1)^2+1) + nu - 0.1) of ITU-R P.526, which is added to the free-space loss. Multiple obstructions are handled by taking the worst single one or by summing them - the Bullington reduction of P.526 is not applied, and the result says which was used. At 30 m the terrain grid is coarse for short paths: a few hundred metres may contain only a handful of samples and the diffraction loss will look jumpy. This suits paths of kilometres or more, and is not a substitute for a surveyed profile.
Common questions
How much does terrain affect a radio link?
Far more than most people expect. Two links the same distance apart, one crossing a ridge and one across flat ground, can differ by 20 dB or more in received level. That is usually the difference between a working link and one that drops calls. The diffraction loss here follows ITU-R P.526, and it grows sharply once the obstruction rises above the line of sight.
What is the Fresnel zone and why does it matter here?
It is the ellipsoidal volume around the direct ray that carries most of the energy. The usual rule is 60% of the first zone clear for free-space behaviour. This tool draws the zone on the profile, shows where the ground enters it, and reports the worst clearance margin, so you can see directly whether the path is cleared or pinched.
How accurate is a 30 m elevation model for this?
It is accurate enough to tell you whether a ridge blocks the link, and too coarse to replace a survey. Over a few hundred metres you may only get a handful of samples and the loss figure will jump around. For a final design, take a measured profile - the geometry here is right, the terrain resolution is the limit.
Should I use the worst obstruction or the sum?
The worst single obstruction corresponds to the Bullington-style treatment in spirit and is the usual planning assumption. Summing is more conservative and suits a path with several comparable obstructions close together. Both are offered, and the result always states which one produced it.
Does this replace a propagation simulation?
No. It is a geometry and clearance check on real terrain, which is what catches the commonest failure - a site placed where a hill is in the way. Capacity, interference and multi-cell behaviour still need a proper planning tool such as Atoll.