The Bulford posts in 3D

Open this to understand this post: https://timdaw37.github.io/bulford-posts-3d/

Drag to rotate, right-drag to pan, scroll to zoom. Click a pit for its name and coordinates.

This summer Phil Harding, Matt Leivers and Fabio Silva published, in PAST 113, a pair of Late Neolithic post-pits at Bulford: 8647 to the south-west and 9019 to the north-east, about 120 m apart. They read the line as 48.1° / 228.1°, a degree off the Stonehenge solstitial axis of 49.1° / 229.1° at around 3000 BC, and they discuss a tangent construction from a claimed base-station pit behind 8647.

I wanted to see the posts on the actual ground, with the actual skyline, and to be able to move the sun.

So here is a 3D plan of the published site. All 39 pits from the Wessex plan on page 3 of the article are marked. The two alignment posts stand in their holes — 8647 at 2 m, 9019 at 4 m, which is how the paper draws them. The rest of the pits are cuts in the chalk, not discs floating on the grass. You can tilt under the landscape and look up through them.

The ground is OS Terrain 50. Switch the drape between hillshade, OpenStreetMap and aerial photography. The red line is the 8647–9019 axis, extended both ways.

The Terrain 50 map with the red post-line, Stonehenge, Woodhenge, Sidbury Hill, and the site inset.

The sun on the skyline

Solstitial sunrise is not a single number. It depends on the height of the local horizon (about 1.61° of azimuth for every degree of altitude at this latitude), on refraction, and on which bit of the disc you call “the sun”. Bulford and Stonehenge are 5 km apart and they do not share a skyline. Using Stonehenge’s axis as the benchmark for Bulford therefore imports Stonehenge’s horizon into a test of Bulford’s geometry.

The viewer puts the rising (or setting) sun on Bulford’s own Terrain 50 horizon. Three buttons choose the convention:

First gleam — upper limb on the skyline
Half orb — centre of the disc
Full orb — lower limb, disc clear

Dates are Midsummer, Midwinter, and the two equinoxes. Epochs are 2950 BC (obliquity 24.02°, close to the date of the posts), 2500 BC, and 2026.

A checkbox draws the sun ray to 8647, which is the observer the paper uses.

From the plan, the posts themselves are at 48.12° grid, 48.31° true (grid convergence at the site is +0.193°). The 3D ground is OS grid; the sun is plotted on that grid, not on a true-north overlay, so the two should sit on the same picture.

On midsummer morning, 2950 BC, first gleam on this horizon comes up at about 48.88°. That is the closest of the three conventions to the posts. Half orb is about 49.27°, full orb about 49.71°. The published 1° gap is largely what you get if you compare a level-horizon 48.1° at Bulford with a horizon-corrected 49.1° at Stonehenge. Like-for-like, first gleam at the two sites is 48.88° against 49.22° — a third of a degree, not a degree.

That is not a demolition of the paper. It is a reason to look. The NE skyline that actually matters is a low ridge at 5.7 km, about +0.42°, not Sidbury Hill. The SW horizon, for the reverse sight, stands a little higher. Whether that is enough for one straight line to serve both midsummer sunrise and midwinter sunset is a question the model is there to argue with: switch to sunset, keep the date, and see.

360° skyline from 8647, with the post azimuth and the quoted Stonehenge 49.1° marked.

A screenshot of the live viewer looking along the posts toward the midsummer sun, first gleam, 2950 BC.

What it is not

The pit positions are digitised from the published 1:1000 site plan, not from the Wessex GNSS archive. That archive has not been released. Two independent readings of 8647 and 9019 agree to 0.2 m, which is good enough for a landscape argument and not good enough to pretend these are survey-grade coordinates. Click a pit and you get easting, northing, an SU grid ref, and lat/lon, with that caveat on the card.

The posts in the holes are schematic. Depths are a working 0.8 m for the two alignment pits and 0.55 m for the others; the paper does not publish pit-base levels. Vertical exaggeration and post-height boost are sliders, because a 4 m timber on a 16 km Terrain 50 sheet is otherwise a pin.

The horizon is Terrain 50, 50 m cells, with earth curvature and Bennett’s altitude-dependent refraction. Environment Agency 1 m LiDAR was not available for the skyline when this was built. A finer DEM would move the numbers; it would not invent a hill that is not there.

Have a go

https://timdaw37.github.io/bulford-posts-3d/

Start with Along posts, Midsummer, 2950 BC, first gleam. Then change one thing at a time: half orb, full orb, 2500 BC, today, sunset. Click 8647 and 9019. Uncheck “All pits” if the scatter is in the way.

The code and the data are at github.com/TimDaw37/bulford-posts-3d.

Harding, P., Leivers, M. and Silva, F. 2026. A newly discovered solstitial post alignment in the Stonehenge landscape at Bulford. PAST 113, 2–5.

Terrain: OS Terrain 50 © Crown copyright, Ordnance Survey (Open Government Licence). Map: © OpenStreetMap contributors. Aerial: ESRI World Imagery. Drafting with Grok; the errors are mine.