St Paul's dome viewshed

A dark ground would have meant nothing here: the question is about sight, not night. So the ground is grey and vermilion is the only strong colour. With nothing else competing, the river reads at once as the one big shape, and that is the finding.

I planned a callout for a long view along a street. The data refused it: the longest land patch that sees the top is 385 m. I used the fall-off with distance instead, which the data do show.

My first draft thickened every visible cell and made the speckle of streets look like the main story. Drawing coverage per pixel instead left the river as the story, which is what the numbers say.

I took the Environment Agency's 5 m height data for the surface as surveyed, put an observer 1.6 m above it and drew a line of sight to the top of the dome from every cell within 5.9 km. The target is my own measurement from the data, 1 m below the highest surface in the 1 m LIDAR near the dome (119.0 m above Ordnance Datum), not an official height. Ground means anything within 2.5 m of bare earth, so streets, parks, squares and water. Rays are sampled every 2.5 m, with a correction for curvature and refraction.

I would take the map outside: check a handful of the predicted views on the ground, and see how it reads on a real phone in daylight rather than a desktop render.