THE ALMANAC — FIELD NOTES · 31°17′ N · 105°51′ W
Notes kept against the turning sky.
A standing record from Carrizal Playa: the alignments STELE I is built to register, two essays on how the sky keeps time, and dispatches from a build reckoned in generations rather than seasons.
[ 01 ] · ALIGNMENT REGISTER · UTC
Upcoming alignments
This is the register STELE I is built to keep: the sun's four turning-points, and the eclipses that fall within reach of the playa, set down to the minute. For now most are observed in the open air, by eye and by survey instrument — only the December solstice will, once the aperture is cut, be registered by the slab itself.
Times are given in Coordinated Universal Time. Sunrise azimuths and noon altitudes are computed for the site at 31°17′ N and are geometric, measured from true north; the real horizon on the lakebed sits a few arcminutes lower, and refraction lifts the disc a touch further.
| UTC | Event | Detail | At Carrizal | STELE I |
|---|---|---|---|---|
| 2026-08-12 · 17:46 | Total solar eclipse | Umbra: Greenland–Iceland–Spain; mag. 1.039 | Below the path — not visible | Crew travels; not registered |
| 2026-08-28 · 04:13 | Partial lunar eclipse | Umbral mag. 0.93; Moon in Aquarius | Visible — Moon high, SE | Logged by eye |
| 2026-09-23 · 00:05 | September equinox | Sun decl. 0°00′; day = night | Sunrise az. 90.0°, due east | Equinox channel — partial cast |
| 2026-12-21 · 20:50 | December solstice | Sun decl. −23°26′; obliquity extreme | Sunrise az. 117.8°; noon alt. 35.3° | Amber beam — the true tick |
| 2027-02-06 · 15:59 | Annular solar eclipse | Antumbra: Pacific–Chile–Atlantic; mag. 0.928 | Below the path — not visible | Not registered |
| 2027-03-20 · 20:24 | March equinox | Sun decl. 0°00′; day = night | Sunrise az. 90.0°, due east | Equinox channel — partial cast |
| 2027-06-21 · 14:10 | June solstice | Sun decl. +23°26′; obliquity extreme | Sunrise az. 62.2°; noon alt. 82.1° | High-sun limit; no beam |
| 2027-08-02 · 10:07 | Total solar eclipse | Umbra: S. Spain–N. Africa–Arabia; max 6m 23s | Below the path — not visible | Crew travels; not registered |
[ 02 ] · ESSAY · ON THE SIDEREAL DAY
How the sky keeps time
A wristwatch keeps the sun's time, and the sun keeps poor time. The Earth turns once on its axis and, in the same span, swings a little further along its orbit, so the sun must be chased a few extra degrees before it returns to the meridian. The figure we call a day — twenty-four hours, flat — is that chase rounded off and averaged across a year. It is a civil convenience, not a measurement.
The stars are more exact. Set against a fixed star instead of the sun, one full turn of the Earth takes 23 hours, 56 minutes, and 4.0905 seconds — the sidereal day. The sky therefore arrives at any given mark 3 minutes 55.91 seconds earlier by the clock each night. A star centred in the polar shaft of STELE I at midnight will find that same centre tomorrow before midnight, and the night after, earlier again. The instrument is tuned to this, not to the sun.
The slip is not error; it is the year, paid out in small change. Across one orbit the stars complete exactly one turn more than the sun: 366.2422 sidereal rotations to 365.2422 sunrises. The extra turn is the orbit itself, written into the spin.
To read the instrument is to read that difference between sun-time and star-time. Four times a year the sun reaches a corner of its travel and seems to pause; once a year, at the December solstice, it stands low enough to enter the aperture and strike the engraved mark. That strike is the sun's single honest report to a star-tuned instrument — one tick, kept once a year, for as long as the concrete stands.
| Quantity | Value |
|---|---|
| Mean solar day | 24h 00m 00.000s |
| Sidereal day | 23h 56m 04.0905s |
| Daily slip (sidereal − solar) | −3m 55.9095s |
| Solar (tropical) year | 365.2422 days |
| Sidereal rotations per year | 366.2422 |
STELE I is set to the stars, not to the sun. The shaft frames the one point that does not move, and lets the rest of the sky turn around it.
[ 03 ] · ESSAY · ON DEEP TIME
The pole star is on loan
Obliquity is the angle of the Earth's lean: 23.44 degrees, or 23°26′, off the vertical to its orbit. It is the whole cause of the seasons — the reason the December sun rides low and the June sun rides high — and it is the figure the studio is named for. It is also, plainly, the slant of the slab. The form leans because the planet leans; the instrument is a working model of the geometry it measures.
That lean is not a constant. Over roughly 41,000 years the obliquity breathes between about 22.1° and 24.5°, and at present it is closing by some 47 arcseconds a century. The consequence is small and certain: each century the solstice sun rises a fraction further along the horizon, and the amber mark cut for our era will, in deep time, fall a hair off true. The instrument is designed to record that drift, not to deny it. A second mark will be needed one day. It will be cut by people we will never meet.
Slower still is precession. The Earth's axis does not hold its aim; it swings in a great cone, tracing a circle of 23.44° radius about the pole of the orbit once every 25,772 years — a little over 50 arcseconds a year. Polaris sits near the pole now by an accident of timing, and only for a while. The polar shaft is cut to hold the true pole, the fixed centre; through it, the stars themselves are seen to move.
This is the measurement the build is for, and it is honest to say that none of us will read it. The pour is reckoned in years and the structure in a century; the precession is reckoned in geological remainder. We are setting a needle into the ground and trusting the sky to pass slowly across it for tens of thousands of years.
| Epoch | Star nearest the pole | Min. separation | Note |
|---|---|---|---|
| c. 2787 BCE | Thuban (α Draconis) | ≈ 0°06′ | Pole star of the pyramid age |
| 1 CE | — (no bright star near) | — | Pole adrift between Kochab and Polaris |
| 2026 CE | Polaris (α Ursae Minoris) | 0°39′ | Present; still closing |
| c. 2100 CE | Polaris | 0°27′ | Closest approach |
| c. 4000 CE | Errai (γ Cephei) | ≈ 3° | Polaris well in retreat |
| c. 7500 CE | Alderamin (α Cephei) | ≈ 3° | Cepheus carries the pole for millennia |
| c. 13700 CE | Vega (α Lyrae) | ≈ 5° | Brightest of all the pole stars |
| c. 27800 CE | Polaris again | — | One full cycle returned |
Polaris will drift from the shaft. The instrument will go on reading — the drifting is the reading.
[ 04 ] · DISPATCH · CARRIZAL PLAYA · JUN 2026
First season on the playa
CARRIZAL PLAYA, 31°17′ N — The lakebed is flat to a degree that unsettles the eye; you can set a level on the cracked clay almost anywhere and it will sit true. We chose this ground for exactly that — a horizon with no argument in it, a sky that goes fully dark, and the long sightlines a polar instrument needs. By late June the surface reads 47°C at midday. Work starts before first light and stops by ten.
The opening weeks were not building but aiming. True north here is not compass north — magnetic declination on the playa runs about 6° east — so the axis was fixed with a gyrotheodolite and then checked, night after night, against the transit of Polaris across the meridian. We hold the line now to better than two arcseconds. Everything the slab will ever measure depends on that one number being right before a single form is raised.
Phase I is the foundation mat: 280 cubic metres of low-heat concrete, 28 percent fly-ash, placed in one continuous night pour with chilled water and embedded cooling loops to keep the core below 70°C as it set. A mass that size makes its own weather inside itself; rush it and it cracks from within, years later. The foundation is poured to outlast the slab.
What stands today is a foundation and a survived first summer — no aperture, no mark, no beam. That is the truth of a hundred-year build at month four: most of it is still a plan and a hole in the desert. Next season raises the first lift of the leaning form and sets the steel sleeve for the polar shaft. We are in no hurry. The instrument is the opposite of a hurry.
- 01Phase
- I of VII — siting & foundation, begun spring 2026.
- 02True north
- Fixed to ±1.8″ by gyrotheodolite; verified on Polaris transits across the meridian.
- 03Foundation mat
- 280 m³ low-heat concrete, 28% fly-ash; core held below 70°C through a continuous night pour.
- 04Design rake
- 31.3° from horizontal — the site latitude, aimed at the celestial pole.
- 05Polar shaft
- 60 mm bore × 6.40 m, steel-sleeved; frames roughly half a degree of sky.
- 06Topping-out
- Projected 2031; instrument calibration to continue for decades after.
[ 05 ] · PLATES · LARGE FORMAT
The plates
We photograph the build on sheet film — 8×10 black-and-white, with a handful of colour plates kept only for the beam. A deep-time instrument deserves a deep-time record, and a glass-backed negative in a cold archive will be legible long after any drive or file format has gone dark. Each plate is exposed slowly and logged like a survey reading: stock, stop, duration, and the sky it was made under.
The plates are not illustrations of the work; they are part of the measurement. A six-hour exposure up the bore turns the night into a set of concentric arcs and shows, in a single frame, the very thing the shaft is built to hold still.
Six-hour exposure up the bore of the survey mock-shaft; star trails close into arcs around the pole.
PLATE II First lightRaking dawn on the west face, Day 41, minutes after sunrise, the form still in rough lift.
PLATE III The rehearsalWinter-solstice geometry modelled with a survey lamp before the aperture is cut, proving where the amber mark must fall.
PLATE IV Moonset on the lakebedThe slab against the dry playa under a setting gibbous moon; four minutes, the clay lit like pewter.
PLATE V The crewSurvey team at the gyrotheodolite, blue hour — the record that hands, not only the sky, made this.
[ 06 ] · THE DISPATCH · QUARTERLY
The Dispatch
The Almanac is updated when the sky earns it. Four times a year — on the two equinoxes and the two solstices — we send The Dispatch: a short field letter from Carrizal Playa carrying the season's alignment register, the latest plates, the survey figures, and a plain account of what the build has done and what it has learned.
It is printed and posted, and also sent by email. There is nothing to buy here and nothing sold; the cadence belongs to the calendar, not to us. If you would keep time with the instrument, leave an address.