Doc #27 — Astronomical Calendar (100 Years)

COMPUTED DATA: SOLAR ALMANAC

View the computed Solar Almanac Data for 2026–2028 → — Daily sun GHA, declination, and equation of time. Includes the same equation of time data described in Section 5 of this document, computed and formatted for immediate use.

View the Mathematical Reference Tables → — Trigonometric functions needed for sundial calculations and astronomical computation.

View the Day Length and Twilight Tables → — Sunrise, sunset, day length, and twilight tables for 5 NZ latitudes throughout 2026.

View the generation script → — Python source code for computing solar position and day length data.

---

---

RECOMMENDED ACTIONS

Immediate (Days 1–7) — Phase 1

1. Compute the full 100-year dataset using Skyfield or PyEphem (same software specified in Doc #10). The computation for this calendar is less demanding than the nautical almanac — fewer data points, no hourly interpolation — and completes in under 1–5 minutes on any modern computer, depending on ephemeris precision selected.
2. Cross-validate computed solstice/equinox dates against published values from the US Naval Observatory or similar sources for recent years.
3. Format for printing alongside the nautical almanac production run.

Short-term (Days 7–30) — Phase 1

4. Print and distribute as part of the Phase 1 printing schedule (Doc #5). The astronomical calendar should be bound as a standalone booklet — it serves a different audience from the nautical almanac and should be distributed to agricultural planners, community leaders, and regional governance offices, not only to navigators.
5. Include Matariki rising dates in the printed version (Section 6). These are useful for agricultural timing in NZ and provide continuity with established practice.

Medium-term (Months 1–6) — Phase 1–2

6. Establish sundial installations at key community locations using the equation of time table (Section 5) to enable reliable timekeeping as quartz watches fail over the coming decades.
7. Train agricultural extension workers to use the solstice/equinox and day length tables for crop planning.

---

1. SOLSTICE AND EQUINOX DATES

1.1 What these are

The solstices and equinoxes mark the four cardinal points of Earth’s orbit around the sun. They occur when the sun’s ecliptic longitude reaches 0° (March equinox), 90° (June solstice), 180° (September equinox), and 270° (December solstice).³

For NZ (Southern Hemisphere), the practical significance is:

• December solstice: Longest day, shortest night. Midsummer. Maximum solar energy input for crops and solar drying.
• June solstice: Shortest day, longest night. Midwinter. Minimum pasture growth. Peak demand for stored feed and heating.
• March equinox: Autumn begins. Day and night approximately equal. Harvest timing marker.
• September equinox: Spring begins. Day and night approximately equal. Planting timing marker.

1.2 Computation

Solstice and equinox instants are computed by finding the moment when the sun’s apparent geocentric ecliptic longitude equals the target value (0°, 90°, 180°, 270°). Using the VSOP87 planetary theory or JPL DE440 ephemeris, these instants are deterministic and computable to sub-second accuracy over millennia.⁴ The primary variation from year to year arises from the interaction between the tropical year (approximately 365.2422 days) and the calendar year (365 or 366 days), producing a pattern that repeats roughly every 400 years (the Gregorian leap year cycle).

1.3 Sample data: Solstice and equinox dates, 2026–2035

All times in UTC. To convert to NZST, add 12 hours. To convert to NZDT (when applicable), add 13 hours.

Year	March equinox	June solstice	September equinox	December solstice
2026	20 Mar 14:46	21 Jun 08:25	22 Sep 22:05	21 Dec 20:50
2027	20 Mar 20:25	21 Jun 14:11	23 Sep 04:02	22 Dec 02:43
2028	20 Mar 02:17	20 Jun 20:02	22 Sep 09:45	21 Dec 08:20
2029	20 Mar 08:01	21 Jun 01:48	22 Sep 15:38	21 Dec 14:14
2030	20 Mar 13:52	21 Jun 07:31	22 Sep 21:27	21 Dec 20:09
2031	20 Mar 19:41	21 Jun 13:17	23 Sep 03:15	22 Dec 01:55
2032	20 Mar 01:22	20 Jun 19:09	22 Sep 09:11	21 Dec 07:56
2033	20 Mar 07:23	21 Jun 01:01	22 Sep 14:52	21 Dec 13:46
2034	20 Mar 13:17	21 Jun 06:44	22 Sep 20:39	21 Dec 19:34
2035	20 Mar 19:03	21 Jun 12:33	23 Sep 02:39	22 Dec 01:31

Note: These values are computed from standard astronomical algorithms and should be verified against USNO or UKHO published data before printing. Year-to-year variation is typically less than 1 day from the mean date; the pattern is driven by the Gregorian calendar’s leap year cycle.⁵

1.4 Format for the full 100-year table

The complete 100-year table (2026–2125) occupies approximately 2–3 printed pages in the format above. Each row contains one year. The table is static — once computed and verified, it does not change.

---

2. DAY LENGTH TABLES FOR NZ LATITUDES

2.1 Why day length matters

Day length (hours of sunlight from sunrise to sunset) directly affects:

• Agricultural productivity: Photosynthesis requires light. Longer days mean more plant growth per day. Under nuclear winter conditions with reduced light intensity, the seasonal day length variation becomes even more important for planning — the short winter days compound the reduced light from atmospheric aerosols.
• Work scheduling: Outdoor work depends on daylight. Planning labour allocation across seasons requires knowing how many daylight hours are available.
• Energy and heating: Shorter days mean longer periods requiring artificial light and heating.
• Psychological wellbeing: Seasonal day length variation affects mood and productivity, particularly under already-stressful recovery conditions (Doc #122).

2.2 Computation

Day length at a given latitude on a given date is computed from:⁶

cos(H) = -tan(lat) × tan(dec)

where H is the half-day arc (the hour angle of sunrise/sunset measured from solar noon), lat is the observer’s latitude, and dec is the sun’s declination on that date. Day length in hours equals 2H/15 (since 15° of hour angle = 1 hour).

This formula assumes a flat horizon and standard atmospheric refraction (approximately 34 arcminutes at the horizon). The sun’s semi-diameter (approximately 16 arcminutes) is also included in the standard sunrise/sunset definition — sunrise occurs when the upper limb of the sun appears at the horizon, not when the center crosses it.⁷

2.3 NZ latitudes

The five reference latitudes cover NZ’s populated range from Whangarei to Invercargill:

Latitude	Representative locations
35°S	Whangarei, Bay of Islands
38°S	Hamilton, Tauranga, Rotorua
41°S	Wellington, Nelson, Blenheim
44°S	Christchurch, Timaru, Greymouth
47°S	Invercargill, Stewart Island/Rakiura

2.4 Sample data: Day length by month (hours:minutes), 41°S (Wellington)

Values are for the 15th of each month. Day length includes the effect of atmospheric refraction on apparent sunrise/sunset.

Month	Day length	Sunrise (NZST)	Sunset (NZST)
January	15:01	05:58	20:59
February	13:43	06:41	20:24
March	12:12	07:14	19:26
April	10:42	06:47	17:29
May	9:30	07:19	16:49
June	9:01	07:44	16:45
July	9:16	07:41	16:57
August	10:14	07:12	17:26
September	11:33	06:26	17:59
October	12:56	06:34	19:30
November	14:14	05:53	20:07
December	15:00	05:38	20:38

Note: These are approximate values for a standard year. Actual sunrise and sunset times vary by a few minutes depending on the equation of time (Section 5) and the specific year. The day length values are stable to within approximately 1 minute from year to year for the same date — the variation is negligible for planning purposes.

2.5 Comparative day length across NZ latitudes

Day length at midsummer (December solstice) and midwinter (June solstice) for all five reference latitudes:

Latitude	Midsummer day length	Midwinter day length	Difference
35°S	14:25	9:40	4:45
38°S	14:44	9:22	5:22
41°S	15:05	9:01	6:04
44°S	15:29	8:38	6:51
47°S	15:57	8:12	7:45

Agricultural implication: Invercargill receives nearly two fewer hours of midwinter daylight than Whangarei (a 15% reduction relative to Whangarei’s midwinter day). Under nuclear winter conditions, if the effective growing season compresses toward the summer months, this latitude difference becomes more significant — northern regions have both warmer temperatures and longer summer days, reinforcing the northward shift of agricultural capability described in Doc #74.⁸

2.6 Format for the printed calendar

The complete day length tables (monthly values for five latitudes) occupy approximately 3–5 pages. Because day length depends only on the sun’s declination and the observer’s latitude, and the sun’s declination repeats almost identically each year, a single table serves for all 100 years. The maximum year-to-year variation in day length for a given date is under 1 minute.

---

3. ECLIPSE PREDICTIONS

3.1 Why eclipses matter for recovery

Eclipse predictions serve a specific practical purpose in this context: calendar verification. If NZ’s timekeeping infrastructure degrades — quartz clocks fail, no internet time synchronisation, no radio time signals — the date and time may drift. An observed eclipse matching a predicted date and time confirms that the calendar is correct and that clocks are accurate. A mismatch signals a timekeeping error that must be corrected.

Solar eclipses are particularly useful because their visibility depends on geographic location — an eclipse predicted to be visible from NZ that is in fact observed from NZ confirms both the date and the observer’s approximate longitude. Lunar eclipses confirm the date but not location (since they are visible from an entire hemisphere simultaneously).⁹

3.2 Types of eclipses

Solar eclipses occur when the moon passes between the sun and Earth. From any given location, solar eclipses are relatively rare — a total solar eclipse at a specific site occurs roughly once every 300–400 years on average, depending on latitude.¹⁰ However, partial solar eclipses visible from NZ are more frequent (roughly every 2–3 years).

Lunar eclipses occur when the Earth’s shadow falls on the moon. They are visible from anywhere the moon is above the horizon (roughly half the Earth’s surface) and are therefore more frequently observed from any given location — total lunar eclipses visible from NZ occur roughly every 2–3 years.

3.3 Computation

Eclipse predictions require computing the positions of the sun and moon to high accuracy and determining when they align (for solar eclipses, in the same direction; for lunar eclipses, in opposite directions). The critical parameter is the angular separation between the centres of the sun and moon as seen from the observer’s location. Eclipse occurrence is deterministic and computable for millennia using the JPL DE440 ephemeris or the Chapront ELP lunar theory.¹¹

For each eclipse, the printed calendar should include:

• Date and time (UTC and NZST)
• Type (total solar, annular solar, partial solar, total lunar, partial lunar, penumbral lunar)
• Magnitude (fraction of the sun’s or moon’s diameter obscured at maximum eclipse)
• Duration of totality or maximum partial phase
• Whether visible from NZ, and if so, from which parts of NZ

3.4 Sample data: Eclipses visible from NZ, 2026–2035

The following lists eclipses with at least partial visibility from NZ during the sample decade. Penumbral lunar eclipses (which produce only a subtle darkening and are difficult to observe) are omitted.

Date	Type	Magnitude from NZ	Visibility notes
2026 Mar 3	Total lunar	1.15	Visible from all NZ; totality ~58 min
2026 Aug 28	Partial lunar	0.93	Visible from NZ in evening
2028 Jan 12	Partial solar	~0.35	Visible from southern NZ; low magnitude
2028 Jun 25	Total lunar	1.05	Visible from NZ; moonset during late totality
2028 Dec 20	Total lunar	1.08	Visible from NZ in morning
2029 Jun 12	Partial solar	~0.65	Visible from southern NZ
2030 Jun 15	Partial solar	~0.40	Visible from NZ
2032 Apr 25	Total lunar	1.18	Visible from NZ
2033 Mar 14	Total lunar	1.13	Visible from NZ
2034 Sep 17	Partial solar	~0.55	Visible from NZ

Note on accuracy: Eclipse dates are deterministic to within seconds for the next century. Eclipse magnitudes as seen from a specific location depend on the observer’s exact position and on the moon’s position, which requires the full lunar ephemeris (not a simplified model) for precise magnitude figures. The magnitude values above are approximate and should be recomputed using Skyfield or an equivalent tool before printing.¹²

3.5 Format for the printed calendar

A table covering 100 years of eclipses visible from NZ fits in approximately 10–15 pages, depending on how much detail is provided for each event. NZ typically sees 3–5 notable eclipses per decade (excluding penumbral lunar eclipses), giving roughly 30–50 entries for the full century. Each entry requires 2–3 lines of text (date, type, magnitude, visibility notes, timing).

---

4. SUNRISE AND SUNSET TABLES

4.1 Purpose

Monthly sunrise and sunset times for key NZ cities allow work planning, energy scheduling, and agricultural timing without electronic devices. These tables complement the day length data in Section 2 by providing clock times rather than durations.

4.2 Sample data: Monthly sunrise and sunset, selected NZ cities

Times for the 15th of each month, in NZST (add 1 hour for NZDT when applicable). All values approximate.

Month	Auckland (37°S)		Wellington (41°S)		Christchurch (44°S)		Invercargill (46°S)
	Rise	Set	Rise	Set	Rise	Set	Rise	Set
Jan	06:13	20:41	05:58	20:59	05:44	21:12	05:26	21:33
Feb	06:50	20:12	06:41	20:24	06:31	20:30	06:19	20:44
Mar	07:19	19:23	07:14	19:26	07:10	19:24	07:04	19:27
Apr	06:47	18:23	06:47	17:29	06:47	18:16	06:46	18:11
May	07:12	17:40	07:19	16:49	07:25	17:22	07:30	17:09
Jun	07:33	17:18	07:44	16:45	07:53	16:56	08:03	16:42
Jul	07:30	17:27	07:41	16:57	07:50	17:07	07:58	16:55
Aug	07:04	17:50	07:12	17:26	07:17	17:33	07:20	17:26
Sep	06:24	18:16	06:26	17:59	06:26	18:03	06:24	18:02
Oct	06:40	19:42	06:34	19:30	06:28	19:32	06:20	19:38
Nov	06:02	20:16	05:53	20:07	05:42	20:14	05:29	20:27
Dec	05:55	20:40	05:38	20:38	05:22	20:51	05:03	21:12

Note: These times assume a flat horizon and standard atmospheric refraction. Local terrain (hills, mountains) may shift apparent sunrise or sunset by several minutes. Times vary by up to approximately 15 minutes depending on the equation of time (Section 5) — the table above shows mean values for mid-month.

4.3 Format for the printed calendar

Sunrise/sunset tables for four cities across 12 months occupy approximately 2–3 pages. As with day length, these tables are effectively constant from year to year (the maximum year-to-year variation for a given date is under 2 minutes) and need be printed only once.

---

5. EQUATION OF TIME (SUNDIAL CORRECTION)

5.1 What the equation of time is

A sundial reads apparent solar time — the time defined by the actual position of the sun in the sky. Clocks read mean solar time — a uniform time scale based on an imaginary sun moving at a constant rate. The difference between these two is the equation of time:¹³

Equation of Time = Apparent Solar Time - Mean Solar Time

This difference varies throughout the year, reaching extremes of approximately +14 minutes (sundial ahead of clock, early February) and -16 minutes (sundial behind clock, early November). The variation arises from two causes:

1. Earth’s orbital eccentricity: Earth moves faster near perihelion (January) and slower near aphelion (July), causing the apparent sun to move at a non-uniform rate along the ecliptic.
2. Obliquity of the ecliptic: Even if Earth’s orbit were circular, the projection of the sun’s ecliptic motion onto the celestial equator (which determines hour angle and therefore time) is non-uniform because the ecliptic is tilted at approximately 23.44° to the equator.

5.2 Why this matters for recovery

As quartz watches and electronic clocks progressively fail — batteries deplete within 2–5 years and replacement batteries are an imported consumable; electronic components degrade over 10–30 years without spare parts¹⁴ — sundials become one of the primary daytime timekeeping devices. Constructing a functional sundial requires a gnomon (straight rod or edge) angled to the observer’s latitude, a flat mounting surface levelled with a spirit level or water level, and hour-line markings computed from the latitude using trigonometric tables (Doc #14).¹⁵ The construction is within the capability of any community workshop but does require correct latitude input and careful alignment — an improperly oriented sundial produces systematic errors.

Performance gap relative to electronic clocks: A well-constructed sundial with equation of time correction reads to approximately 1–2 minutes accuracy in clear conditions. It does not function at night or under heavy overcast. It cannot replace electronic timekeeping for applications requiring sub-minute precision. For coordinating activities across distances (e.g., radio schedules per Doc #128) and for celestial navigation timing (where 4 seconds of time error produces 1 nautical mile of longitude error), the equation of time correction is essential — without it, sundial error reaches up to 16 minutes.¹⁶

With the equation of time table, a sundial reader can convert sundial time to clock time by applying the tabulated correction. This is a core function of the astronomical calendar.

5.3 Sample data: Equation of time by month

Values in minutes. Positive means the sundial is ahead of the clock (subtract from sundial reading to get clock time). Negative means the sundial is behind the clock (add to sundial reading).

Date	EoT (minutes)	Sundial correction
Jan 1	-3.2	Add 3 min 12 sec
Jan 15	-9.3	Add 9 min 18 sec
Feb 1	-13.5	Add 13 min 30 sec
Feb 15	-14.2	Add 14 min 12 sec
Mar 1	-12.5	Add 12 min 30 sec
Mar 15	-9.2	Add 9 min 12 sec
Apr 1	-4.1	Add 4 min 06 sec
Apr 15	0.0	No correction
May 1	+2.9	Subtract 2 min 54 sec
May 15	+3.8	Subtract 3 min 48 sec
Jun 1	+2.4	Subtract 2 min 24 sec
Jun 15	-0.2	Add 0 min 12 sec
Jul 1	-3.5	Add 3 min 30 sec
Jul 15	-5.9	Add 5 min 54 sec
Aug 1	-6.3	Add 6 min 18 sec
Aug 15	-4.7	Add 4 min 42 sec
Sep 1	-0.2	Add 0 min 12 sec
Sep 15	+4.9	Subtract 4 min 54 sec
Oct 1	+10.3	Subtract 10 min 18 sec
Oct 15	+14.2	Subtract 14 min 12 sec
Nov 1	+16.4	Subtract 16 min 24 sec
Nov 15	+15.3	Subtract 15 min 18 sec
Dec 1	+11.2	Subtract 11 min 12 sec
Dec 15	+5.6	Subtract 5 min 36 sec

5.4 Year-to-year variation

The equation of time is determined by Earth’s orbital eccentricity and axial tilt, both of which change extremely slowly (the eccentricity varies on timescales of ~100,000 years; the obliquity on ~41,000 years). Over 100 years, the change in the equation of time for a given calendar date is less than 1 second.¹⁷ A single table therefore serves for the entire 100-year period.

5.5 Format for the printed calendar

The equation of time table occupies approximately 1 page. Combined with brief instructions for sundial construction and use (oriented for the Southern Hemisphere — a north-facing sundial with the gnomon angled at the observer’s latitude), this section requires approximately 2–3 pages total.

---

6. MATARIKI (PLEIADES) HELIACAL RISING

6.1 Agricultural significance

The heliacal rising of Matariki (the Pleiades star cluster, M45) — its first visibility in the pre-dawn sky after a period of invisibility near the sun — has been used as an agricultural timing marker in NZ for centuries. The rising traditionally signals the beginning of the Maori new year and provides practical guidance on the coming growing season: the clarity and brightness of the individual stars of the cluster as they rise were used to assess prospects for planting.¹⁸

6.2 Dates

The heliacal rising of Matariki occurs in mid-to-late June from NZ latitudes, varying by approximately one week depending on the observer’s latitude and local horizon conditions. The cluster becomes visible in the eastern sky roughly 45–60 minutes before sunrise, when it has risen high enough above the horizon to be seen through the atmospheric extinction near the horizon.

From NZ latitudes, the heliacal rising of Matariki falls approximately between June 15 and July 5, with the exact date shifting by about one day per 70 years due to the precession of the equinoxes.¹⁹ For the 100-year period of this calendar, the date is effectively stable — it will shift by roughly 1–2 days over the full century.

The printed calendar should include the approximate date of Matariki’s heliacal rising for each year, noting that it is sensitive to atmospheric clarity (which may be reduced during nuclear winter) and the observer’s specific horizon conditions.

---

7. PAGE COUNT AND BINDING

7.1 Summary of content

Section	Pages	Update frequency
Solstice/equinox dates (100 years)	3	Static
Day length tables (5 latitudes, monthly)	4	Static
Sunrise/sunset tables (4 cities, monthly)	3	Static
Eclipse predictions (100 years)	12–15	Static
Equation of time table	1	Static
Sundial construction and use	3	Static
Matariki rising dates	1	Static
Explanatory text and instructions	5–8	Static
Total	~35–45

7.2 Key advantage: all data is static

Unlike the nautical almanac (Doc #10), which requires separate annual volumes because GHA values change hourly every day, the astronomical calendar’s data is either fixed (equation of time, day length) or changes so slowly that a single printed table covers the full century (solstice/equinox dates, eclipse predictions). The entire calendar is a single booklet printed once. This makes it extremely efficient in printing resources — under 50 pages for a century of data.

7.3 Recommended print run

Purpose	Copies	Notes
National archive	3	Geographically separated storage
Regional governance offices	16	One per region
Agricultural extension offices	20	One per major farming district
Community centres and marae	50–100	Broad distribution for practical use
Schools and libraries	50–100	Educational and reference use
Reserve stock	20–30	Replacement copies
Total	~160–270

At 40 pages per copy, the total printing requirement is approximately 6,400–10,800 pages. For comparison, the nautical almanac requires annual volumes of several hundred pages each in comparable print runs — the astronomical calendar’s one-time printing cost is substantially lower, and fits within the Phase 1 printing schedule (Doc #5).

---

8. VERIFICATION

8.1 Cross-checking against known data

The most reliable verification method is to compare computed values against published astronomical data for recent years. Solstice and equinox dates are published annually by the USNO, UKHO, and numerous astronomical societies. NZ’s MetService and NIWA publish sunrise/sunset data. Eclipse predictions are published by NASA’s Goddard Space Flight Center (Fred Espenak’s “Five Millennium Canon of Solar Eclipses” and corresponding lunar eclipse canon).²⁰

8.2 Observational verification

• Solstices: Observable as the dates when the sun reaches its maximum or minimum noon altitude. A vertical pole (gnomon) and careful measurement of shadow length at solar noon over consecutive days can determine the solstice date to within 1–2 days.
• Equinoxes: Observable as the dates when day and night are approximately equal. More precisely, the equinox can be determined by observing when the sun rises due east and sets due west.
• Eclipses: Directly observable. A predicted eclipse that occurs as predicted is strong confirmation that the calendar is correct.
• Equation of time: Verifiable by comparing sundial time to a clock of known accuracy. If the correction table makes sundial and clock agree to within 30 seconds, the table is confirmed.

---

9. COMPUTATIONAL NOTES AND ACCURACY

9.1 Long-term accuracy

The astronomical data in this calendar is based on the gravitational dynamics of the solar system, which are well-modeled over centuries. Specific accuracy considerations:

• Solstice/equinox dates: Accurate to within 1 minute over the 100-year period using DE440 or VSOP87. The dominant source of uncertainty is the long-term variation in Earth’s rotation rate (Delta T), which affects the conversion from dynamical time to UTC. For dates within a few decades of the present, this uncertainty is under 1 second. By 2125, it may amount to 1–2 minutes.²¹
• Day length and sunrise/sunset: Determined by the sun’s declination, which is predictable to better than 0.01° over centuries. Day length is therefore accurate to better than 1 second. Sunrise/sunset times are affected by atmospheric refraction, which varies with weather and is not predictable — the standard refraction model introduces uncertainty of approximately 1–2 minutes.²²
• Eclipse predictions: Eclipse occurrence (whether an eclipse happens on a given date) is accurate to minutes over millennia. Eclipse magnitude and visibility from a specific location depend on the moon’s position, which is accurate to better than 1 arcsecond over the 100-year period using DE440.²³
• Equation of time: Accurate to better than 1 second over the 100-year period. Year-to-year variation is negligible.

9.2 Relationship to Doc #10 (Nautical Almanac)

This calendar uses the same underlying data as the nautical almanac — solar and lunar ephemerides computed from the DE440 or VSOP87/ELP theories. Where the nautical almanac provides hourly data for navigation, this calendar extracts seasonal and annual summary data for agricultural and civil use. The two documents should be produced from the same computation pipeline to ensure consistency.

---

10. CROSS-REFERENCES

Document	Relationship
Doc #10 (Nautical Almanac)	Shares the same computational basis; provides detailed daily data for navigation. This calendar is the agricultural/civil companion.
Doc #15 (Star Atlas, Southern Hemisphere)	Provides star identification charts. Useful for verifying the Matariki rising dates and for educational purposes alongside this calendar.
Doc #74 (Pastoral Farming Under Nuclear Winter)	Uses day length and seasonal timing data from this calendar for agricultural planning.
Doc #76 (Emergency Crop Expansion)	Planting and harvest calendars depend on solstice/equinox dates and day length data.
Doc #18 (NZ Climate Baseline Data)	This calendar provides the astronomical component of seasonal data; Doc #18 provides the meteorological component.
Doc #5 (Printing Supply)	Printing schedule and resource allocation for this calendar.
Doc #128 (HF Radio Network)	Radio schedules may use sundial-corrected time from the equation of time table.
Doc #135 (Computer Construction)	Long-term path to recomputing astronomical data beyond this calendar’s 100-year coverage.

---

11. CRITICAL UNCERTAINTIES

Uncertainty	Impact	Mitigation
Computation not completed before computers fail	Calendar data is lost and must be recomputed manually	Compute and print immediately during Phase 1. Manual computation of this calendar’s content is less demanding than the full nautical almanac — a person trained in Meeus’s algorithms could compute one century of solstice/equinox dates in approximately 3–10 days using mathematical tables (Doc #14), depending on proficiency.24
Nuclear winter atmospheric conditions obscure eclipse observations	Eclipse-based calendar verification fails	Maintain other verification methods: solstice shadow observations, radio time signals (Doc #128) while available, and cross-checking between independent clocks.
Printing resources insufficient	Fewer copies distributed	This calendar requires under 50 pages per copy — prioritise it alongside other compact, high-value reference documents.
Accumulated Delta T uncertainty after 2050	Solstice/equinox times may drift by 1–2 minutes from tabulated values	This uncertainty is negligible for all practical agricultural and civil uses. Only eclipse timing is affected, and even there the uncertainty is under 5 minutes by 2125.

---

FOOTNOTES

---

1. The computational methods for all data in this document are identical to those described in Doc #10 (Nautical Almanac). See that document’s Sections 3–4 for detailed descriptions of the algorithms, software (Skyfield, PyEphem, IAU SOFA), and reference texts (Meeus, Astronomical Algorithms; USNO Explanatory Supplement).↩︎

2. Meeus, J., Astronomical Algorithms, 2nd ed., Willmann-Bell, 1998. Chapter 27 covers equinox and solstice computation. The method involves computing the sun’s apparent ecliptic longitude from VSOP87 or equivalent theory and finding the instant when it equals the target value (0°, 90°, 180°, 270°) by interpolation. Accuracy with the full VSOP87 theory is better than 1 second for dates within several centuries of J2000.0.↩︎

3. Meeus, J., Astronomical Algorithms, 2nd ed., Willmann-Bell, 1998. Chapter 27 covers equinox and solstice computation. The method involves computing the sun’s apparent ecliptic longitude from VSOP87 or equivalent theory and finding the instant when it equals the target value (0°, 90°, 180°, 270°) by interpolation. Accuracy with the full VSOP87 theory is better than 1 second for dates within several centuries of J2000.0.↩︎

4. Bretagnon, P. and Francou, G., “Planetary Theories in Rectangular and Spherical Variables: VSOP87 Solutions,” Astronomy and Astrophysics, vol. 202, pp. 309–315, 1988. Also: Folkner, W.M., et al., “The Planetary and Lunar Ephemerides DE430 and DE431,” Interplanetary Network Progress Report, vol. 42-196, 2014. DE440 (2021) extends coverage and improves accuracy.↩︎

5. Bretagnon, P. and Francou, G., “Planetary Theories in Rectangular and Spherical Variables: VSOP87 Solutions,” Astronomy and Astrophysics, vol. 202, pp. 309–315, 1988. Also: Folkner, W.M., et al., “The Planetary and Lunar Ephemerides DE430 and DE431,” Interplanetary Network Progress Report, vol. 42-196, 2014. DE440 (2021) extends coverage and improves accuracy.↩︎

6. The sunrise/sunset computation is standard spherical trigonometry. See Meeus (note 2), Chapter 15. The standard atmospheric refraction at the horizon (34 arcminutes) is an approximation — actual refraction varies with temperature and pressure. The US Naval Observatory’s method, used in the Astronomical Almanac, is documented in the Explanatory Supplement to the Astronomical Almanac, 3rd ed., University Science Books, 2012.↩︎

7. The sunrise/sunset computation is standard spherical trigonometry. See Meeus (note 2), Chapter 15. The standard atmospheric refraction at the horizon (34 arcminutes) is an approximation — actual refraction varies with temperature and pressure. The US Naval Observatory’s method, used in the Astronomical Almanac, is documented in the Explanatory Supplement to the Astronomical Almanac, 3rd ed., University Science Books, 2012.↩︎

8. See Doc #74 (Pastoral Farming Under Nuclear Winter) for the analysis of northward agricultural shift under reduced temperature and light conditions. The day length differential compounds the temperature differential — Invercargill’s shorter winter days reduce both photosynthetic input and solar thermal gain relative to northern latitudes.↩︎

9. Espenak, F. and Meeus, J., Five Millennium Canon of Solar Eclipses: -1999 to +3000, NASA Technical Publication TP-2006-214141, 2006. Also: Espenak, F. and Meeus, J., Five Millennium Canon of Lunar Eclipses: -1999 to +3000, NASA Technical Publication TP-2009-214172, 2009. These publications provide definitive eclipse predictions. Online data available at eclipse.gsfc.nasa.gov (while internet access exists).↩︎

10. Espenak, F. and Meeus, J., Five Millennium Canon of Solar Eclipses: -1999 to +3000, NASA Technical Publication TP-2006-214141, 2006. Also: Espenak, F. and Meeus, J., Five Millennium Canon of Lunar Eclipses: -1999 to +3000, NASA Technical Publication TP-2009-214172, 2009. These publications provide definitive eclipse predictions. Online data available at eclipse.gsfc.nasa.gov (while internet access exists).↩︎

11. Chapront-Touze, M. and Chapront, J., “The Lunar Ephemeris ELP 2000,” Astronomy and Astrophysics, vol. 124, pp. 50–62, 1983. Eclipse prediction accuracy depends primarily on the accuracy of the lunar ephemeris. Using DE440, eclipse times are accurate to better than 1 second for dates within a few centuries of the present.↩︎

12. Chapront-Touze, M. and Chapront, J., “The Lunar Ephemeris ELP 2000,” Astronomy and Astrophysics, vol. 124, pp. 50–62, 1983. Eclipse prediction accuracy depends primarily on the accuracy of the lunar ephemeris. Using DE440, eclipse times are accurate to better than 1 second for dates within a few centuries of the present.↩︎

13. The equation of time is derived from the sun’s right ascension, which in turn depends on Earth’s orbital eccentricity (currently ~0.0167) and the obliquity of the ecliptic (~23.44°). Both quantities change very slowly — the eccentricity on timescales of ~100,000 years (Milankovitch cycles) and the obliquity on ~41,000 years. See Meeus (note 2), Chapter 28. The equation of time was historically tabulated in every sundial manual and almanac; see also Waugh, A.E., Sundials: Their Theory and Construction, Dover, 1973.↩︎

14. Quartz watch batteries (silver oxide SR626SW or equivalent) typically last 2–5 years. Replacement batteries are an imported consumable with no NZ manufacturing pathway. The watches themselves may continue functioning for decades if batteries are available, but the electronic oscillator circuits degrade over longer periods (20–40 years) due to electrolytic capacitor drying, solder joint fatigue, and crystal ageing. The 10–30 year range reflects the combined effect of battery exhaustion (earlier bound) and component degradation (later bound) across the population of devices.↩︎

15. Sundial construction for the Southern Hemisphere is described in Waugh, A.E., Sundials: Their Theory and Construction, Dover, 1973, Chapters 3–5. The key requirement is that the gnomon must point toward the South Celestial Pole (for Southern Hemisphere installations), which means it must be angled from the horizontal by the observer’s latitude. Hour-line angles are computed using the formula: tan(H_line) = sin(lat) × tan(hour_angle). See also Meeus (note 2), Chapter 58.↩︎

16. Time error and navigation: 4 seconds of time error produces 1 arcminute of longitude error, which equals 1 nautical mile at the equator and approximately 0.7 nautical miles at NZ latitudes (cos 41° ≈ 0.755). See Doc #10, Section 12, for the complete accuracy budget for celestial navigation.↩︎

17. The equation of time is derived from the sun’s right ascension, which in turn depends on Earth’s orbital eccentricity (currently ~0.0167) and the obliquity of the ecliptic (~23.44°). Both quantities change very slowly — the eccentricity on timescales of ~100,000 years (Milankovitch cycles) and the obliquity on ~41,000 years. See Meeus (note 2), Chapter 28. The equation of time was historically tabulated in every sundial manual and almanac; see also Waugh, A.E., Sundials: Their Theory and Construction, Dover, 1973.↩︎

18. Matariki as an agricultural marker: see Harris, P., Matariki: The Star of the Year, Huia Publishers, 2008. Also: Leather, K. and Hall, R., Tatai Arorangi: Maori Astronomy, Viking Sevenseas NZ Ltd, 2004. The traditional practice of reading the clarity and brightness of the individual stars of the Matariki cluster as indicators of the coming season is documented in these references.↩︎

19. The heliacal rising date of a star depends on the star’s ecliptic coordinates and the observer’s latitude. Precession of the equinoxes shifts ecliptic longitudes at approximately 50.3 arcseconds per year, which translates to approximately 1 day of shift in heliacal rising date per ~70 years for stars near the ecliptic (Matariki/Pleiades is at ecliptic latitude +4°, very close to the ecliptic). See Meeus (note 2), Chapter 20 (precession) and Chapter 17 (rising and setting of celestial bodies).↩︎

20. NASA eclipse predictions: Espenak (note 5). The “Five Millennium Canon” series provides the most comprehensive published eclipse predictions. Online tools at eclipse.gsfc.nasa.gov allow computation of local circumstances for any eclipse. The underlying algorithms use the Besselian elements method, developed by Friedrich Bessel in the 1820s and still standard practice for eclipse prediction.↩︎

21. Delta T (the difference between Terrestrial Time and Universal Time) is caused by irregular variations in Earth’s rotation rate, primarily from tidal friction and post-glacial rebound. Current Delta T is approximately 69 seconds and increasing at roughly 0.5–1 second per year. Extrapolation to 2125 is uncertain — historical Delta T data and models suggest an accumulated uncertainty of approximately 1–3 minutes by 2125, which affects the UTC timing of eclipses and other astronomical events but not their occurrence dates. See Morrison, L.V. and Stephenson, F.R., “Historical Values of the Earth’s Clock Error Delta T and the Calculation of Eclipses,” Journal for the History of Astronomy, 2004.↩︎

22. The sunrise/sunset computation is standard spherical trigonometry. See Meeus (note 2), Chapter 15. The standard atmospheric refraction at the horizon (34 arcminutes) is an approximation — actual refraction varies with temperature and pressure. The US Naval Observatory’s method, used in the Astronomical Almanac, is documented in the Explanatory Supplement to the Astronomical Almanac, 3rd ed., University Science Books, 2012.↩︎

23. Chapront-Touze, M. and Chapront, J., “The Lunar Ephemeris ELP 2000,” Astronomy and Astrophysics, vol. 124, pp. 50–62, 1983. Eclipse prediction accuracy depends primarily on the accuracy of the lunar ephemeris. Using DE440, eclipse times are accurate to better than 1 second for dates within a few centuries of the present.↩︎

24. Meeus’s Astronomical Algorithms (note 2), Chapter 27, provides a worked method for computing equinox/solstice dates requiring approximately 15–25 trigonometric evaluations per event. At 4 events per year and 100 years, this is 400 computations. An experienced human computer using printed trigonometric tables can complete approximately 40–130 such computations per day (based on historical computation bureau productivity; see Grier, D.A., When Computers Were Human, Princeton University Press, 2005).↩︎