Recoverable Foundation

Recovery Library

Doc #172 — Long-Term Archival Strategy

Preserving Critical Knowledge Across Decades and Centuries in New Zealand

Phase: 3+ (Years 3–100+, with preparatory actions in Phase 1–2) | Feasibility: [A] Established

Unreliable — not for operational use. Produced by AI under human direction and editorial review. This document contains errors of fact, judgment, and emphasis and has not been peer-reviewed. See About the Recovery Library for methodology and limitations. © 2026 Recoverable Foundation. Licensed under CC BY-ND 4.0. This disclaimer must be included in any reproduction or redistribution.

EXECUTIVE SUMMARY

Digital storage degrades within years without maintenance; printed documents on standard acidic paper become brittle within decades; the electronics required to read digital media are themselves depleting assets.1 Without a deliberate archival strategy, the knowledge that made recovery possible is lost to NZ’s descendants — and with it, the ability to maintain the industrial, agricultural, and medical capabilities that this entire library exists to preserve.

Every other document in the Recovery Library addresses a problem with a time horizon of months, years, or at most a few decades. This document addresses the longer question: how does NZ ensure that the knowledge accumulated during recovery — and the pre-war knowledge captured while digital systems still function — survives for 50, 100, or 500 years? Laser-printed documents on standard office paper last 20–100 years under controlled conditions, but most commercial paper manufactured since the mid-19th century uses alum-rosin sizing and will become brittle within decades.2 The Recovery Library itself, and the vast body of technical reference data NZ needs (Docs #10–28), will physically decay unless action is taken.

The honest assessment: NZ can produce archival-quality paper from domestic materials (Doc #29), and properly stored acid-free paper lasts centuries. Microfilm requires silver halide chemistry that NZ may not sustain domestically, but existing stocks provide a bridge of years to decades. For the most critical information, inscription on stone or fired ceramic is proven to last millennia and is within NZ’s capability. The binding constraint is not technology — it is institutional will: maintaining archival programs across generations when immediate survival demands compete for resources.

Contents

Phase 1 (Months 0–12):

  1. Print the entire Recovery Library and all precomputed reference data (Docs #10–28) on the best available paper at full quality (Doc #5, #30)
  2. Inventory all NZ microfilm equipment, unexposed film stocks, and processing chemicals — a one-time asset that cannot be replaced
  3. Secure all existing archival collections (National Library, Archives NZ, museum collections, iwi archives)
  4. Begin digital-to-print conversion of the highest-priority digital-only knowledge (Section 3)

Phase 2 (Years 1–3):

  1. Begin microfilming the most critical printed documents while equipment and film stock remain
  2. Establish acid-free paper production trials at Kinleith or Kawerau mills (Doc #29)
  3. Train archival conservation staff — paper repair, binding, environmental management
  4. Identify suitable archival storage sites across NZ (Section 2)

Phase 3+ (Years 3–100+):

  1. Establish routine acid-free paper production for archival documents
  2. Begin stone and ceramic inscription of the most critical knowledge (Section 1.4)
  3. Distribute archival copies to regional repositories and marae
  4. Establish the institutional framework for multi-generational maintenance (Section 5)

ECONOMIC JUSTIFICATION

Immediate costs (Phase 1–2): ~18–51 person-years total

Component Person-years Notes
Printing Recovery Library and reference data 2–5 Part of existing printing program (Doc #5, #29)
Microfilm inventory and operation 3–8 Requires trained operators
Digital-to-print conversion 10–30 Competing with other printing demands
Archival site preparation 2–5 Surveying, shelving, environmental monitoring
Conservation training 1–3 10–20 people in basic archival conservation

Ongoing costs (Phase 3+): ~14–35 person-years per year

Acid-free paper production (1–3), stone/ceramic inscription workshop (5–15), repository maintenance at 5–10 sites (5–10), conservation and repair (2–5), administration (1–2).

Value: The archival program insures against catastrophic knowledge loss. If accumulated technical knowledge — metallurgical processes, agricultural methods, medical procedures, engineering specifications — is destroyed in year 50 or 100, it must be relearned from scratch at a cost of decades. The annual cost (14–35 person-years) is less than 0.002% of the national workforce. NZ currently employs approximately 600 staff at the National Library and Archives NZ combined.3 The proposed program is smaller than the pre-war archival establishment.

1. MEDIA DURABILITY

1.1 The durability hierarchy

Medium Expected lifespan NZ producibility
Stone inscription (granite, basalt) 1,000–10,000+ years [A] Suitable stone and masons available
Fired ceramic (stoneware) 1,000–10,000+ years [A] Suitable clays and kilns available
Microfilm (silver halide, polyester base) 500+ years [C] Cannot produce; existing stocks only
Acid-free paper (alkaline-sized) 300–1,000+ years [B] Producible with effort
Steel plate (engraved or etched) 500+ years in shelter (stainless); 100–300 years (mild steel, coated) [B] Mild steel from NZ Steel; stainless from repurposed stock only
Standard laser print on office paper 20–100 years [A] While toner lasts
Digital (hard drive / SSD) 3–10 years without power [A] Short-term only
Digital (optical disc) 5–50 years [A] Short-term; reader availability declines
Digital (magnetic tape, LTO) 15–30 years [A] Medium-term; drive availability the constraint

There is an inverse relationship between information density and durability. Stone holds the least information per kilogram but lasts the longest. Digital media hold the most but last the shortest time. Paper occupies the middle ground: moderate density, moderate durability, producible from NZ materials.45678

1.2 Paper: the workhorse medium

Most paper manufactured since approximately 1850 uses alum-rosin sizing, which generates sulfuric acid as it ages, breaking down cellulose fibres until the paper crumbles. Standard commercial paper is acidic — documents printed on it will deteriorate within 20–50 years and may be unreadable within 100.9

Acid-free paper, produced using alkaline sizing (calcium carbonate instead of alum-rosin), meeting ISO 9706, lasts several hundred years.10 Rag paper — cotton or linen fibres rather than wood pulp — is even more durable; 15th-century examples remain in good condition after 500+ years.11

NZ capability: The Kinleith and Kawerau mills produce chemical pulp from radiata pine (Doc #29). Converting to alkaline sizing requires calcium carbonate (available from NZ limestone, particularly Golden Bay Cement’s Tarakohe quarry; Doc #97), reformulated sizing agents, and retention aid chemistry adjustments. The mills must also modify their wet-end chemistry — replacing alum-rosin with alkyl ketene dimer (AKD) or alkenyl succinic anhydride (ASA) as sizing agents. The dependency chain for AKD/ASA: both are derived from fatty acids (AKD from fatty acid chlorides via ketene dimerisation; ASA from maleic anhydride and linear internal olefins). Fatty acid feedstock is available in NZ from tallow (Doc #85) and animal fats; maleic anhydride is a petrochemical not domestically producible. AKD synthesis from tallow-derived fatty acids is therefore the more realistic domestic pathway, but requires acyl chloride chemistry (thionyl chloride or phosphorus trichloride, both requiring chemical manufacturing capability). The realistic Phase 2–3 pathway is to use existing imported AKD/ASA stocks while developing domestic fatty-acid sizing chemistry; if those stocks are exhausted before domestic production is established, a transitional alkaline sizing using starch-based or rosin-lime systems may be required, with lower permanence than full AKD/ASA sizing but better than acidic paper. The resulting paper will be acid-free but will likely have lower tear resistance and fold endurance than specialised archival stock meeting ISO 9706, because radiata pine kraft pulp produces shorter fibres than the cotton or linen rag used in premium archival paper.12 Timeline: 2–5 years from Phase 1.

NZ also has wool and harakeke fibre (Doc #100) as potential rag paper feedstock. Harakeke fibre is longer and stronger than wood pulp but produces a coarser sheet than cotton or linen rag — acceptable for archival durability but less suitable for fine-print reproduction.13 Rag paper production is a hand process requiring fibre preparation (retting, beating), vat forming, couching, pressing, and drying — labour-intensive at approximately 20–50 sheets per worker per day, but feasible from Phase 3 onward for limited quantities of premium archival stock.

1.3 Microfilm

Properly processed silver halide microfilm on polyester base has a life expectancy exceeding 500 years at 18–20°C, 30–40% RH.14 A single 35mm reel (100 feet) holds approximately 600–2,500 pages depending on reduction ratio and source document size; at the standard 24x reduction for A4 pages, approximately 600–800 pages per reel is realistic.15 The entire Recovery Library, estimated at approximately 25,000–35,000 pages (based on an average of 15–40 pages per document across 172 documents — actual page count requires physical audit), would require approximately 30–60 reels at standard reduction ratios. Microfilm can be read without electricity — a significant advantage over digital media — but requires a quality optical reader with 10–25x magnification and precise focus; a standard hand magnifying glass (typically 2–5x) is insufficient to resolve 24x-reduced text legibly.16 Dedicated microfilm readers (light table + precision optics) are the practical minimum; improvised readers using camera lenses or telescope eyepieces are possible but significantly harder to use.

NZ has microfilm equipment at the National Library (Wellington), Archives NZ (Wellington and regional offices), and university libraries including the University of Auckland, University of Canterbury, and University of Otago. Unexposed film stock is finite and non-renewable: silver halide emulsion production requires silver nitrate (from metallic silver — NZ has limited silver mining at Coromandel/Hauraki), gelatin (from animal hides — available domestically), halide salts, and photographic-grade processing chemicals including hydroquinone-based developer and sodium thiosulfate fixer.17 The window for microfilm production is probably 5–15 years, depending on the quantity of unexposed stock discovered during Phase 1 inventory. After that, existing reels become read-only assets — durable for centuries but irreplaceable.

1.4 Stone and ceramic: millennial durability

For information that must survive regardless of institutional continuity, stone and ceramic inscription are the only proven technologies. Mesopotamian cuneiform tablets have survived 4,000–5,000 years. The Rosetta Stone dates to 196 BCE.18

Stone: NZ has Coromandel granite, Central Otago schist, various basalts, and Oamaru limestone.19 A skilled mason inscribes approximately 16–48 words per day in granite — working at roughly 10–30 letters per hour in hard stone, which at five characters per word over an eight-hour day yields that range — slow, but the information is effectively permanent.20 Pounamu (greenstone) is extremely durable but culturally significant to Māori (taonga) and should not be used without iwi agreement.

Fired ceramic: Stoneware fired to 1,200°C+ matches stone in durability. NZ has suitable ball clays (Northland, Waikato) and fire clays, with both electric and wood-fired kilns available.21 Text impressed into unfired clay and kiln-fired can be produced at several hundred words per day — faster than stone carving, and ceramic tablets can be mass-produced using stamps or moulds. The disadvantage: ceramic breaks on impact and requires careful handling.

Metal plates: Engraved or acid-etched mild steel offers approximately 5–10x more information per unit area than stone inscription (finer line width achievable), is very durable if kept dry and coated, and is producible from NZ Steel’s Glenbrook works output; stainless steel is not a standard Glenbrook product and requires chromium and nickel not mined domestically.22 The full dependency chain for acid-etched plates: (1) NZ Steel produces flat-rolled mild steel from Waikato iron sands; (2) plates must be cut to size and surface-finished (requiring abrasives and metal-working tools); (3) a resist must be applied — bitumen or beeswax work but must be sourced domestically; (4) etchant must be produced — ferric chloride requires iron, hydrochloric acid (from salt + sulfuric acid, Doc #113), and an oxidiser; nitric acid requires the Ostwald process or fuming acid stocks; (5) spent etchant must be neutralised and disposed of safely. Ferric chloride is the more practicable option if hydrochloric acid production is established; nitric acid etching requires more hazardous chemistry. Acid etching is faster than hand engraving (approximately 500–2,000 words per plate per day vs. 24–64 words per day for hand engraving, based on 15–40 characters per hour by a skilled engraver at 5 characters per word over 8 hours)23 but produces shallower marks that may be harder to read after centuries of surface corrosion. Mild steel plates require protective coatings — lacquer, wax, or dry sealed storage — to prevent corrosion; without protection, surface rust will obscure shallow-etched text within decades in NZ’s humid coastal environment.

1.5 Digital media: a bridge, not an archive

Digital storage is a working medium, not an archival one. Hard drives fail mechanically. SSDs lose data through charge leakage within 1–10 years without power, depending on temperature and NAND technology (TLC/QLC cells degrade faster than SLC).24 Optical discs vary wildly in quality. Even intact media become inaccessible as the electronics to read them fail (Doc #129). Digital systems should be used aggressively in Phase 1–2 to compile and distribute knowledge, then that knowledge must be committed to durable physical media. The AI facility (Doc #129) is the most important knowledge generator in this period, and its outputs must be preserved in physical form.

2. STORAGE ENVIRONMENTS

The primary enemies of paper are heat, moisture, light, pests, and atmospheric pollutants. The lifespan difference between optimal and poor storage is large: ISO 11799 accelerated aging studies indicate that paper lasting several centuries in controlled archival conditions may degrade to an unusable state within 30–50 years in uncontrolled warm, humid, and light-exposed storage — a roughly 5–10x reduction in expected lifespan.25

Target conditions: Paper: 15–20°C, 30–50% RH, dark, clean air, pest-free. Microfilm: 18–20°C, 30–40% RH, sealed containers.2627

Best NZ regions: Central Otago and inland Canterbury (cool, dry, 400–600mm rainfall), Manawatu and Wairarapa hill country, southern Hawke’s Bay. Avoid: coastal locations (humidity, salt), Wellington (extreme seismic risk), West Coast South Island (3,000–6,000mm rainfall), low-lying areas (flood risk), geothermal zones (hydrogen sulfide corrodes paper and metal).28

Low-technology solutions: Earth-sheltered storage exploits NZ’s stable subsurface temperature (~12–15°C at 2–5m depth).29 Humidity control requires drainage and ventilation. Central Otago wine cellars demonstrate that suitable underground environments exist. Purpose-built concrete-lined vaults are achievable (Doc #97). Thick-walled masonry buildings (walls 400mm+) provide thermal stability without active climate control — the Vatican Secret Archives, established in the 17th century, and the Bodleian Library (Oxford, 1602) demonstrate that thick masonry construction maintains acceptable archival conditions over centuries.30

3. WHAT TO PRESERVE: PRIORITISATION

3.1 Priority tiers

Tier 1 — Stone/ceramic inscription candidates (must survive institutional collapse):

Fundamental science (periodic table, basic physics, chemistry, anatomy). Key mathematical tables (Doc #14). Agricultural fundamentals calibrated for NZ. Metallurgy. The existence and location of other repositories. A historical summary of what happened and NZ’s situation — context for future readers.

Tier 2 — Acid-free paper, multiple distributed copies:

The full Recovery Library (172 documents). Precomputed reference data (Docs #10–28). Medical knowledge. Engineering specifications. Agricultural research. Heritage skills documentation (Doc #160). Legal and governance frameworks, including the Treaty of Waitangi. Te reo Māori resources and mātauranga Māori (iwi-governed collections — Section 3.3).

Tier 3 — Standard paper; microfilm where possible:

Broader scientific literature. Historical records. Literature, art, music. Economic and social data. Maps.

3.2 The philosophical question

The instinct is to prioritise practical survival knowledge — how to smelt iron, grow food, treat disease. But practical knowledge is context-dependent and may become obsolete. Fundamental scientific understanding enables future generations to derive new practices for conditions we cannot predict. A farming manual tells you what to do; a soil science textbook tells you why and enables adaptation. For long-term preservation, the “why” arguably matters more than the “what.”

Historical knowledge — understanding what happened and why — provides context that enables future generations to make sense of their situation. A generation that knows it is recovering from a nuclear war makes different decisions from one that does not understand why its civilisation is diminished.

3.3 Iwi-governed archival collections

Archival preservation of Māori knowledge — mātauranga Māori, te reo Māori resources, and iwi-specific historical records — must follow the same governance principles as heritage skills documentation (Doc #160): Māori ownership and Māori governance over content and access. Some knowledge carries access restrictions that must be respected, because violating them destroys the trust needed for Māori participation, resulting in less knowledge preserved, not more. The practical approach: iwi-managed collections at marae or iwi-designated locations, with copies at regional repositories by iwi agreement. The national program provides materials and technical support (acid-free paper, binding, storage environment guidance); iwi govern content and access. This also provides an independent institutional layer for archival redundancy — iwi governance structures have demonstrated multi-generational institutional continuity, making them a natural partner for archival programs that must persist across centuries.

4. REDUNDANCY AND GEOGRAPHIC DISTRIBUTION

A single repository anywhere in NZ is not safe enough for knowledge that must survive centuries. NZ sits on the Pacific Ring of Fire; earthquakes, volcanic eruptions, floods, and fires are real risks over multi-century timescales.31

Minimum 5 copies of Tier 1 and 2 material, geographically separated:

Repository Region Rationale
National Library / Archives NZ Lower North Island Existing infrastructure and expertise
Northern repository Auckland or Waikato Different seismic zone; largest population
Central South Island Canterbury or Otago Dry climate; different seismic zone
Southern repository Southland or Central Otago Maximum geographic separation
Iwi-managed repository Variable Separate institutional governance

Additionally: every regional public library, participating marae, polytechnic campus, and hospital library should hold printed copies of relevant material.

The “bottle” strategy: For Tier 1 knowledge, sealed concrete vaults at 5+ geologically stable inland sites, containing stone or ceramic tablets plus sealed stainless steel tubes with microfilm and acid-free paper. Durable surface markers in English and te reo Māori. This is a backup of last resort, designed to survive even if institutional maintenance ceases for generations.32

5. INSTITUTIONAL FRAMEWORK

5.1 The real problem

The technical challenges are solvable. The hard problem is ensuring archival programs are maintained across decades and centuries, through changes of government, shifts in priorities, and generational turnover. Every generation faces the temptation to defer archive maintenance for more pressing needs.

5.2 Lessons from history

Monastic libraries (500–1500 CE) survived centuries because knowledge preservation was core to the institution’s identity.33 Chinese imperial archives persisted through dynasty changes because bureaucracy depended on them.34 Modern national archives are sustained by legal mandate but vulnerable to budget cuts. The lessons: embed archival mission in institutional identity; make the archive operationally useful; legal mandates help but cultural valuation matters more.

5.3 Recommendations

  1. Legal mandate: Establish the program in legislation (building on the Public Records Act 2005).35
  2. Distributed governance: Spread responsibility across the National Library, Archives NZ, university libraries, iwi archives, and regional libraries. If one fails, others continue.
  3. Operational integration: Make the archive useful for current governance, education, and research — an archive nobody uses is eventually abandoned.
  4. Cultural embedding: Integrate archival preservation into the educational curriculum. Future generations must experience the archive’s usefulness firsthand.
  5. Māori partnership: Iwi-governed collections connect knowledge preservation to living communities with multi-generational institutional memory.
  6. Succession planning: Conservation skills (paper repair, binding, environmental management) must be transmitted to new practitioners. Integrate archival conservation into the polytechnic training system (Doc #157).

6. CRITICAL UNCERTAINTIES

Uncertainty Impact Resolution
NZ microfilm stock quantity Determines how much can be microfilmed Skills census inventory (Doc #8)
Acid-free paper production feasibility Determines long-term paper capability Trial production at Kinleith/Kawerau
Institutional persistence across generations The fundamental risk to the entire program Design for usefulness, not obligation
Seismic/volcanic risk to specific sites Loss of repositories over centuries Geographic redundancy (5+ sites)
Quality of NZ-produced archival paper Determines actual document lifespan Testing, quality control, iterative improvement
Digital-to-print conversion completeness Some digital knowledge will be lost Prioritise ruthlessly (Section 3)
Future generations’ willingness to maintain archives Unknowable Cultural embedding; operational integration

CROSS-REFERENCES

  • Doc #1 — National Emergency Stockpile Strategy (securing archival materials)
  • Doc #5 — Printing Supply Requisition (toner and paper stocks)
  • Doc #8 — National Asset and Skills Census (microfilm inventory; archival workforce)
  • Doc #10–28 — Precomputed Reference Data (primary archival content)
  • Doc #29 — Paper and Ink Production (domestic paper capability)
  • Doc #30–31 — Print Optimisation and Manual Printing Methods
  • Doc #97 — Cement and Concrete (calcium carbonate; vault construction)
  • Doc #100 — Harakeke Fiber Processing (rag paper feedstock)
  • Doc #113 — Sulfuric Acid (chemical dependency for etching and processing)
  • Doc #129 — AI Facility Operations (knowledge generation requiring preservation)
  • Doc #134 — Computing Self-Sufficiency Roadmap (digital media constraints)
  • Doc #144 — Emergency Powers and Democratic Continuity (legal authority for archival mandate)
  • Doc #150 — Treaty of Waitangi and Māori Governance (iwi archival partnership)
  • Doc #157 — Accelerated Trade Training (conservation skills training)
  • Doc #160 — Heritage Skills Preservation and Transmission (documentation requiring preservation; conservation skills; Māori knowledge documentation and partnership protocols, §4.5–4.7)
  • Doc #168 — Recovery Library Master Index (catalogue preservation)

FOOTNOTES

  1. NIST digital data sustainability reports; Rosenthal, D.S.H., “Keeping Bits Safe: How Hard Can It Be?”, Communications of the ACM, 2010. SSDs lose data through charge leakage — JEDEC JESD218A specifies 1 year retention at 40°C for consumer SSDs; lower temperatures extend this but multi-year unpowered storage is unreliable. Hard drive failure rates documented by Backblaze and Google studies. Optical disc durability varies: pressed factory DVDs may last 50+ years; recordable discs using organic dyes may fail within 5–15 years.↩︎

  2. Barrow Research Laboratory findings; Library of Congress preservation reports, 1950s–1990s. Alum-rosin sized paper generates sulfuric acid internally via acid hydrolysis of cellulose. Acidic paper becomes brittle within 50–100 years under average library conditions. Rag paper from the 15th century remains flexible and strong. See also: Cunha, G.M., “Conservation of Library Materials,” Scarecrow Press, 1967.↩︎

  3. National Library of NZ and Archives NZ staffing from publicly available annual reports. National Library ~350–400 staff; Archives NZ ~200–250. Trained conservation specialists are a small fraction — perhaps 20–40 nationally. Figures should be verified against current data.↩︎

  4. NIST digital data sustainability reports; Rosenthal, D.S.H., “Keeping Bits Safe: How Hard Can It Be?”, Communications of the ACM, 2010. SSDs lose data through charge leakage — JEDEC JESD218A specifies 1 year retention at 40°C for consumer SSDs; lower temperatures extend this but multi-year unpowered storage is unreliable. Hard drive failure rates documented by Backblaze and Google studies. Optical disc durability varies: pressed factory DVDs may last 50+ years; recordable discs using organic dyes may fail within 5–15 years.↩︎

  5. Barrow Research Laboratory findings; Library of Congress preservation reports, 1950s–1990s. Alum-rosin sized paper generates sulfuric acid internally via acid hydrolysis of cellulose. Acidic paper becomes brittle within 50–100 years under average library conditions. Rag paper from the 15th century remains flexible and strong. See also: Cunha, G.M., “Conservation of Library Materials,” Scarecrow Press, 1967.↩︎

  6. ISO 18923:2000, “Imaging materials — Polyester-base magnetic tape — Storage practices.” LTO tape rated 15–30 years at 18–22°C, 20–50% RH. LTO formats are backward-compatible only two generations; drive obsolescence makes tapes inaccessible regardless of media condition.↩︎

  7. AIIM microfilm guidelines; ISO 18911:2010, “Imaging materials — Processed safety photographic films — Storage practices.” Silver halide on polyester base assessed at 500+ year life expectancy based on accelerated aging tests (Arrhenius method), confirmed by excellent condition of 90-year-old archival microfilm. Readable with ~10–25x magnifying lens and ambient light.↩︎

  8. Durability of stone and ceramic demonstrated by Mesopotamian cuneiform tablets (4,000–5,000 years), Egyptian stone inscriptions (4,000+ years), Roman inscriptions (2,000 years). Materials are intrinsically stable — many survived burial and abandonment, not careful maintenance. The Rosetta Stone (196 BCE, granodiorite) is one well-known example.↩︎

  9. Barrow Research Laboratory findings; Library of Congress preservation reports, 1950s–1990s. Alum-rosin sized paper generates sulfuric acid internally via acid hydrolysis of cellulose. Acidic paper becomes brittle within 50–100 years under average library conditions. Rag paper from the 15th century remains flexible and strong. See also: Cunha, G.M., “Conservation of Library Materials,” Scarecrow Press, 1967.↩︎

  10. ISO 9706:1994, “Paper for documents — Requirements for permanence.” Requires minimum pH 7.5, alkaline reserve equivalent to 2% calcium carbonate, minimum tear resistance and fold endurance. Expected lifespan “several hundred years” — based on accelerated aging and historical sample extrapolation.↩︎

  11. Barrow Research Laboratory findings; Library of Congress preservation reports, 1950s–1990s. Alum-rosin sized paper generates sulfuric acid internally via acid hydrolysis of cellulose. Acidic paper becomes brittle within 50–100 years under average library conditions. Rag paper from the 15th century remains flexible and strong. See also: Cunha, G.M., “Conservation of Library Materials,” Scarecrow Press, 1967.↩︎

  12. Radiata pine kraft pulp has a weight-weighted average fibre length of approximately 2.5–3.0 mm, compared to 15–25 mm for cotton linter fibres. Shorter fibres produce weaker inter-fibre bonding, reducing tear resistance. See: Kibblewhite, R.P., “Radiata Pine Wood Residue Fibre Characteristics,” Appita Journal, various. ISO 9706 specifies minimum tear resistance (350 mN) and fold endurance (150 double folds); NZ-produced alkaline-sized pine paper may meet these thresholds but with less margin than cotton rag stock.↩︎

  13. Harakeke (Phormium tenax) fibre length is approximately 5–15 mm after muka processing — longer and stronger than wood pulp but shorter than cotton linters. The fibre is coarser (approximately 10–15 µm diameter vs. 8–10 µm for cotton), producing a rougher sheet surface. See: Carr, D.J., et al., “Fibre Properties of Phormium tenax,” Textile Research Journal, 2005.↩︎

  14. AIIM microfilm guidelines; ISO 18911:2010, “Imaging materials — Processed safety photographic films — Storage practices.” Silver halide on polyester base assessed at 500+ year life expectancy based on accelerated aging tests (Arrhenius method), confirmed by excellent condition of 90-year-old archival microfilm. Readable with ~10–25x magnifying lens and ambient light.↩︎

  15. AIIM/ANSI standard MS23 and MS111. At 24x reduction ratio, a standard 100-foot roll of 35mm microfilm holds approximately 600–800 A4 pages in duplex (two pages per frame). Higher reduction ratios (42x or 48x) increase capacity but reduce readability without precision optics. Older references citing 2,400+ pages per reel typically assume 16mm film at high reduction ratios with smaller source documents.↩︎

  16. AIIM microfilm guidelines; ISO 18911:2010, “Imaging materials — Processed safety photographic films — Storage practices.” Silver halide on polyester base assessed at 500+ year life expectancy based on accelerated aging tests (Arrhenius method), confirmed by excellent condition of 90-year-old archival microfilm. Readable with ~10–25x magnifying lens and ambient light.↩︎

  17. Silver halide emulsion requires silver nitrate (from metallic silver — NZ has limited mining at Coromandel/Hauraki; Doc #22), gelatin (from animal hides — available), and halide salts. Processing requires hydroquinone-based developer and thiosulfate fixer. NZ’s domestic chemical capability may be insufficient for photographic-grade production.↩︎

  18. Durability of stone and ceramic demonstrated by Mesopotamian cuneiform tablets (4,000–5,000 years), Egyptian stone inscriptions (4,000+ years), Roman inscriptions (2,000 years). Materials are intrinsically stable — many survived burial and abandonment, not careful maintenance. The Rosetta Stone (196 BCE, granodiorite) is one well-known example.↩︎

  19. GNS Science geological publications. NZ formations include Coromandel granite, Central Otago schist (layered structure less ideal for inscription), various volcanic basalts, and Oamaru limestone (easy to carve but weathers faster than igneous rock).↩︎

  20. Estimated from historical practice and modern letter-cutting. A skilled mason produces approximately 10–30 letters per hour in granite. At ~5 characters per word over an 8-hour day: (10–30) × 8 ÷ 5 = 16–48 words per day. Rates vary significantly with stone hardness, letter size, and mason skill. Softer stones such as Oamaru limestone allow higher rates (possibly 50–100 words/day) but have lower durability. The body text uses 16–48 words/day for granite specifically.↩︎

  21. NZ has ball clays (Northland, Waikato) and fire clays suitable for stoneware. Firing to 1,200–1,300°C achievable in electric or wood kilns. Historical precedent: Mesopotamian ceramic tablets, Chinese ceramic records. See also: Brickell, B., “NZ Pottery.”↩︎

  22. NZ Steel’s Glenbrook works produces flat-rolled steel products from iron sands. Stainless steel production is not standard Glenbrook output — stainless requires chromium and nickel alloying, which NZ does not mine domestically. Mild steel plates are the realistic NZ-produced option, though these require protective coatings or dry storage to prevent corrosion. Existing stainless steel stock (structural, kitchen, dairy equipment) could be repurposed for high-priority inscriptions. Engraving rates estimated from modern practice: V-cut lettering in mild steel at approximately 15–40 characters per hour by hand. At 5 characters/word × 8 hours: (15–40) × 8 ÷ 5 = 24–64 words/day by hand engraving. Acid etching rates depend on plate size and resist application method; approximately 500–2,000 words per plate per production day is achievable if the chemical supply chain (etchant, resist, neutralisation) is established and operating. Body text has been corrected to reflect the 24–64 words/day engraving rate.↩︎

  23. NZ Steel’s Glenbrook works produces flat-rolled steel products from iron sands. Stainless steel production is not standard Glenbrook output — stainless requires chromium and nickel alloying, which NZ does not mine domestically. Mild steel plates are the realistic NZ-produced option, though these require protective coatings or dry storage to prevent corrosion. Existing stainless steel stock (structural, kitchen, dairy equipment) could be repurposed for high-priority inscriptions. Engraving rates estimated from modern practice: V-cut lettering in mild steel at approximately 15–40 characters per hour by hand. At 5 characters/word × 8 hours: (15–40) × 8 ÷ 5 = 24–64 words/day by hand engraving. Acid etching rates depend on plate size and resist application method; approximately 500–2,000 words per plate per production day is achievable if the chemical supply chain (etchant, resist, neutralisation) is established and operating. Body text has been corrected to reflect the 24–64 words/day engraving rate.↩︎

  24. NIST digital data sustainability reports; Rosenthal, D.S.H., “Keeping Bits Safe: How Hard Can It Be?”, Communications of the ACM, 2010. SSDs lose data through charge leakage — JEDEC JESD218A specifies 1 year retention at 40°C for consumer SSDs; lower temperatures extend this but multi-year unpowered storage is unreliable. Hard drive failure rates documented by Backblaze and Google studies. Optical disc durability varies: pressed factory DVDs may last 50+ years; recordable discs using organic dyes may fail within 5–15 years.↩︎

  25. ISO 11799:2015, “Document storage requirements for archive and library materials.” Also: Ogden, S. (ed.), “Preservation of Library & Archival Materials,” NEDCC, 1999. The five environmental factors: temperature, humidity, light, air quality, pest control.↩︎

  26. ISO 11799:2015, “Document storage requirements for archive and library materials.” Also: Ogden, S. (ed.), “Preservation of Library & Archival Materials,” NEDCC, 1999. The five environmental factors: temperature, humidity, light, air quality, pest control.↩︎

  27. ISO 18911:2010 — microfilm storage. Extended-term (500+ years): temperature below 21°C, RH 20–30%. Processing quality (thorough fixer washing) critical — poorly processed microfilm deteriorates within decades.↩︎

  28. Archives NZ has publicly acknowledged seismic vulnerability of its Wellington repository. Seismic strengthening undertaken, but the fundamental risk of housing irreplaceable records in a high-seismic city remains. The 2016 Kaikōura earthquake (M7.8) highlighted this. See Archives NZ annual reports at https://www.archives.govt.nz/.↩︎

  29. Ground temperature below seasonal variation zone (~5–15m depth) approximates local mean annual air temperature. Central Otago: ~10°C. Waikato: ~14°C. Both within archival range. Data from GNS Science. Note: Taupō Volcanic Zone has elevated subsurface temperatures — unsuitable.↩︎

  30. Vatican Apostolic Archive (formerly Secret Archives) has maintained documents since 1612 in thick-walled masonry construction. The Bodleian Library, Oxford (founded 1602) houses manuscripts dating to the 9th century. Both demonstrate centuries-scale preservation through thermal mass without mechanical climate control, though both benefited from the relatively mild, stable English and Italian climates. NZ’s Central Otago and inland Canterbury have comparable temperature stability.↩︎

  31. GNS Science, “National Seismic Hazard Model for New Zealand.” NZ experiences ~15,000 earthquakes per year. Major events (M7+) occur on average every few decades. Volcanic eruptions in the Taupō Volcanic Zone and Taranaki are lower-frequency but high-impact.↩︎

  32. Precedents: Svalbard Global Seed Vault (Norway, 2008) for crop genetic diversity; Memory of Mankind project (Austria) deposits ceramic tablets in a salt mine. These demonstrate interest in ultra-long-term preservation but have not been tested over design timescales.↩︎

  33. Harris, M.H., “History of Libraries in the Western World,” Scarecrow Press, 1995; Battles, M., “Library: An Unquiet History,” Norton, 2003. Monastic libraries sustained knowledge through centuries of political upheaval because the institution was self-supporting and knowledge preservation was core to its identity.↩︎

  34. Chinese imperial archives (mingshi records and the Grand Secretariat archives) spanned multiple dynasties including Han, Tang, Song, Ming, and Qing, with record-keeping continuity maintained across dynastic transitions because the incoming bureaucracy required access to precedent, land records, and administrative history. See: Hucker, C.O., “A Dictionary of Official Titles in Imperial China,” Stanford University Press, 1985; Wilkinson, E., “Chinese History: A Manual,” Harvard-Yenching Institute, 2000. Archival continuity was institutional, not purely cultural — the new government had an administrative interest in preserving the old records.↩︎

  35. Public Records Act 2005 (NZ). https://www.legislation.govt.nz/act/public/2005/0040/late... — Establishes framework for managing public records and archives. Provides a legal foundation extensible to the long-term archival program described here.↩︎

← Back to Library Index