Dev.to WebDev 🛠 Dev 👁 0 📖 3 min read

Computing All 13 New Moons of 2027 in Python & Pure JS (Zero Runtime Dependencies)

In my last post about rebuilding Eastman Kodak's 1928–1989 13-month calendar for 2027, several readers asked about an unexpected astronomical coincidence: why does 2027 have 13 New Moons, and how do you compute lunar ill

In my last post about rebuilding Eastman Kodak's 1928–1989 13-month calendar for 2027, several readers asked about an unexpected astronomical coincidence: why does 2027 have 13 New Moons, and how do you compute lunar illumination accurately in a static browser app without shipping a 500 KB ephemeris library?

Here is the exact pipeline I used for SOL-2027 — combining PyEphem / Jean Meeus UTC tables at build time with a tiny O(1) lookup and cosine phase interpolator in pure JavaScript.

Why Gregorian August 2027 Collides — and a 13×28 Calendar Doesn't

A synodic lunar month averages 29.530588 days, which means a 365-day solar year spans ~12.368 lunar cycles. Most years squeeze 12 New Moons across 12 uneven Gregorian months, occasionally forcing a "Black Moon" (two New Moons in one Gregorian month).

In 2027, the very first New Moon hits early (Jan 7, 20:25 UTC), leaving just enough room for a 13th New Moon on Dec 27, 20:13 UTC. On the Gregorian calendar, August 2027 gets crammed with two New Moons (Aug 2 and Aug 31).

When you project those exact same 13 UTC instants onto Moses Cotsworth's 13-month × 28-day International Fixed Calendar (where Month 07 SOL runs June 18 – July 15 around the Summer Solstice):

  • 01 · January (Jan 1–28): Jan 7 · 20:25 UTC (Day 7)
  • 02 · February (Jan 29–Feb 25): Feb 6 · 15:57 UTC (Day 9)
  • 03 · March (Feb 26–Mar 25): Mar 8 · 09:30 UTC (Day 11)
  • 04 · April (Mar 26–Apr 22): Apr 6 · 23:52 UTC (Day 12)
  • 05 · May (Apr 23–May 20): May 6 · 10:59 UTC (Day 14)
  • 06 · June (May 21–Jun 17): Jun 4 · 19:41 UTC (Day 15)
  • 07 · SOL (Jun 18–Jul 15): Jul 4 · 03:03 UTC (Day 17)
  • 08 · July (Jul 16–Aug 12): Aug 2 · 10:06 UTC (Day 18)
  • 09 · August (Aug 13–Sep 9): Aug 31 · 17:42 UTC (Day 19)
  • 10 · September (Sep 10–Oct 7): Sep 30 · 02:37 UTC (Day 21)
  • 11 · October (Oct 8–Nov 4): Oct 29 · 13:37 UTC (Day 22)
  • 12 · November (Nov 5–Dec 2): Nov 28 · 03:25 UTC (Day 24)
  • 13 · December (Dec 3–Dec 30): Dec 27 · 20:13 UTC (Day 25)

Zero skipped months. Zero double-moon collisions. Each New Moon advances smoothly by ~1.5 days across the 28-day grid from Day 7 in January to Day 25 in December.

Precomputing Exact Brackets for Zero-Dependency Runtime Math

Instead of bundling a heavy orbital mechanics library into the client, I precomputed the 13 exact 2027 UTC New Moon timestamps plus two boundary anchors (Dec 9, 2026 00:52 UTC and Jan 26, 2028 15:12 UTC) so Jan 1–6 and Dec 28–31 never lose edge accuracy:

const PREV_NEW_MOON_2026_MS = Date.UTC(2026, 11, 9, 0, 52);
const NEXT_NEW_MOON_2028_MS = Date.UTC(2028, 0, 26, 15, 12);

const ALL_NEW_MOON_BRACKETS = [
  PREV_NEW_MOON_2026_MS,
  ...NEW_MOONS_2027.map((m) => m.utcMs),
  NEXT_NEW_MOON_2028_MS
];

At runtime, finding the lunar cycle progress and illuminated fraction for any day in 2027 takes ~12 lines of pure JS:

const noonMs = Date.UTC(2027, month - 1, day, 12, 0, 0);
let prevMs = ALL_NEW_MOON_BRACKETS[0];
let nextMs = ALL_NEW_MOON_BRACKETS[1];

for (let i = 0; i < ALL_NEW_MOON_BRACKETS.length - 1; i++) {
  if (noonMs >= ALL_NEW_MOON_BRACKETS[i] && noonMs < ALL_NEW_MOON_BRACKETS[i + 1]) {
    prevMs = ALL_NEW_MOON_BRACKETS[i];
    nextMs = ALL_NEW_MOON_BRACKETS[i + 1];
    break;
  }
}

const cycleProgress = (noonMs - prevMs) / (nextMs - prevMs); // 0 = New, 0.5 = Full
const illumination = Math.round(((1 - Math.cos(2 * Math.PI * cycleProgress)) / 2) * 100);

Because each lunation is bracketed by its actual variable synodic length (which fluctuates between ~29.27 and ~29.83 days due to lunar orbital eccentricity) rather than a naive constant 29.53, this tiny routine stays tightly synced with real New & Full Moon dates all year long while adding < 1 KB gzipped.

Explore the Full 2027 Lunar & Fixed Calendar Tool

If you want to inspect the live UTC table, check what moon phase lands on your 2027 birthday, or see the archival print & gift editions:

Have you ever used precomputed ephemeris brackets instead of runtime astronomy packages in client apps? Would love to hear how you handle timezone/UTC edge cases!

📰 Read the original article on Dev.to WebDev

Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.