LIT_074: Puente-Sierra et al. (2025) — A National Assessment of the Drinking Water Infrastructure Deficit in New…

LIT_074: Puente-Sierra et al. (2025) — A National Assessment of the Drinking Water Infrastructure Deficit in New…

Source

doi:10.1080/13241583.2025.2481694 — original publication (opens in a new tab; the file is not redistributed)

Puente-Sierra et al. (2025) — A National Assessment of the Drinking Water Infrastructure Deficit in New Zealand by Territorial Authority and Sociodemographic Characteristics

First national standardised geospatial assessment of NZ's TA-owned drinking-water reticulation — 57,174 km of pipe, 30.7% in poor/very-poor condition, 18.5% past life-expectancy, mean weighted age 37.8 yr across 66 of 67 TAs and 625 of 629 water distribution zones (4,135,000 people, ~88% of NZ). Retrieved for RT_316 to verify the ~NZ$23bn drinking-water reticulation network value CR_050 rests on.

✅ 23bn VERIFIED PRESENT** (Introduction, p.35, verbatim) — CR_050's attribution to this paper is correct, **NOT a mis-attribution**. ⚠ **But it is a repeated PRIOR estimate**, not the paper's own valuation: the paper derives it as NZ85bn total three-waters value (WICS 2021) × 27% drinking-water-networks share (Water NZ 2022) ≈ 23bn, and its central finding is that this deficit is likely a **substantial underestimate** (its measured 30.7% poor/very-poor is >100% higher than TAs self-reported to Taumata Arowai). ⚠ The **~3,300–5,000/connection is NOT in the paper — that is purely CR_050’s synthesis derivation (RT_316’s own framing confirmed). Treat $23bn as a soft top-down floor, not a hardened per-connection cost.

Summary

This is the first national, standardised, geospatial condition assessment of New Zealand’s Territorial-Authority-owned drinking-water reticulation. The authors (GeoHealth Laboratory, University of Canterbury) collated pipe location, material and installation-age data from 66 of the 67 TAs — standardising 698 raw material values and 262 asset-type values into a single ArcGIS dataset — and graded condition from material-specific life expectancies (low / midpoint / high scenarios). The final dataset covers 57,174 km of pipe (92.2% of the raw data) across 625 of 629 water distribution zones. Nationally, 30.7% of the reticulation is in poor or very poor condition and 18.5% is past its life expectancy (midpoint), with a mean weighted pipe age of 37.8 years; the network is 67.3% plastic and 19.8% asbestos cement. The study finds wide variation between TAs (poor/very-poor 2.4–75.4%), with older cities and some small rural TAs (Kawerau 75.4%, Gisborne 58.7%) in the worst condition, and inequities by area deprivation. Its policy conclusion is that NZ’s reported water infrastructure deficit is likely a substantial underestimate and that TA-based amalgamation (“Local Water Done Well” CCOs) faces real economies-of-scale barriers. For Neobiome Intelligence the load-bearing item is the paper’s Introduction, which states the prior estimate that “drinking water reticulation networks could be worth ~NZ$23bn” — the exact figure CR_050 cited and RT_316 was raised to verify.

Key claims

- claim: "🎯 LOAD-BEARING (RT_316). VERBATIM: 'Water infrastructure (drinking water, wastewater and stormwater) in NZ is believed to be worth around NZ$85bn, and it faces an investment deficit of approximately NZ$120–185bn over the next 30 years (Water Industry Commission for Scotland 2021). Public drinking water networks could be worth as much as 80% of all drinking water assets and 27% of all water assets in NZ (Water NZ 2022). Based on this, drinking water reticulation networks could be worth ~NZ$23bn and require ~NZ$32–50bn of investment over the next 30 years. However, these prior estimates have been based on incomplete and non-standardised data from the 67 Territorial Authorities (TAs).' ⚠ PROVENANCE: the ~NZ$23bn is a PRIOR estimate the paper REPEATS and derives (NZ$85bn × 27% ≈ $23bn), NOT a valuation the paper itself measured; the ultimate primaries are WICS 2021 (the $85bn total + $120–185bn three-waters deficit) and Water NZ 2022 National Performance Review (the 27% / 80% asset-share proportions). The paper's own contribution is the physical CONDITION data (below), and its argument is that these monetary estimates are likely a SUBSTANTIAL UNDERESTIMATE. NO per-connection figure appears anywhere in the paper — the ~$3,300–5,000/connection in CR_050 is a synthesis derivation, not from this source."
  source_location: "1. Introduction, p.35 (paragraph 2)"
- claim: "National reticulation stock + condition (the paper's own measured finding). VERBATIM (Abstract + §3.1 + §4.1): 'In total, 30.7% of the 57,174 km of drinking water pipes in NZ were in poor or very poor condition, while 18.5% were past their life expectancy.' The mean weighted pipe age is 37.8 years (range 20.4–54.3 across TAs). Midpoint condition splits: very poor 21.2% + poor 9.5% = 30.7% PVPC; past life-expectancy 18.5% (low 8.7% – high 28.1%). Total reticulation length ranges from a minimum of 71 km to a maximum of 11,768 km across TAs."
  source_location: "Abstract, p.35; §3.1 Descriptive statistics + Table 1, pp.40–41; §4.1 Infrastructure deficit, p.43"
- claim: "Asset-type composition of the reticulation (the reticulation-vs-service-line split). VERBATIM (§2.3.1.1): pipes were classified as 'transmission or trunk mains (3.0% of our dataset), distribution or rider mains (81.1%), service lines (12.2%), abandoned (2.0%), private (1.0%), other (0.5%) and unknown (0.3%). Abandoned pipes were discarded for further analysis, but private pipes were included when connected to a TA-owned WDZ.' → distribution/rider mains dominate the network (81.1%); service lines are 12.2%."
  source_location: "§2.3.1.1 Asset type, p.36"
- claim: "Coverage / denominator basis. VERBATIM (§2.2): 'These 629 WDZs serve water to 4,135,000 people (~88% of the total population)', across 'the 67 Territorial Authorities (TAs) responsible for the majority of drinking water provision in NZ'. The final analytic dataset (57,174 km = 92.2% of the raw dataset) covered 625 of the 629 TA-owned WDZs and 66 of the 67 TAs; the 4 excluded WDZs supplied <0.01% of the NZ population. (The paper reports population served, NOT number of connections — so it does not itself support a per-connection denominator.)"
  source_location: "§2.2 Population and study area, p.36; §2.3.3 Data exclusions + §3.1, pp.38–40"
- claim: "Pipe material composition (Table 1). Most pipes are plastic (67.3%): polyethylene 36.0%, polyvinyl chloride 31.3%; asbestos cement 19.8%; cast iron 4.8%; steel 4.7%; copper 1.3%; galvanised 1.0%; ductile iron 0.6%; other 0.5%. 46.2% of NZ drinking-water pipes are over 40 years old (comparable to the USA 47.0%, surpassing Japan 20.7%, Spain 26.0%, Canada 41.0%)."
  source_location: "§3.1 + Table 1, p.41; §4.1, p.43 (over-40-year comparison)"
- claim: "Headline argument — the deficit is likely UNDERESTIMATED (bounds how $23bn/$120–185bn should be used). VERBATIM (§4.1): 'On balance, our results suggest that previously reported estimates of the infrastructure deficit for drinking water reticulation may be a substantial underestimate.' And (§4.4): 'Our results suggest the reported water infrastructure deficit of NZ$120–185bn over the next 30 years (Water Industry Commission for Scotland 2021) could be substantially higher. Our estimate on the proportion of pipes in PVPC is more than 100% greater than that reported by TAs to Taumata Arowai.' (The paper's PVPC estimate 30.7% vs 13.0% self-reported by TAs; e.g. Watercare reported 0% while the paper estimated 40% of its 12,000 km network was PVPC.)"
  source_location: "§4.1 Infrastructure deficit, p.43; §4.4 Policy implications, p.44"
- claim: "Small rural TAs — the remote-community bind (directly relevant to NI scope). VERBATIM (§4.2): 'More problematic are some smaller rural TAs that have a substantial proportion of their pipes in suboptimal condition and small ratepayer bases such as Kawerau (75.4% in PVPC, population on public supply 7,100), Gisborne (58.7%, 35000), South Wairarapa (57.5%, 7,100) Gore (54.2%, 10200), South Waikato (52.4%, 19400) or South Taranaki (51.0%, 21900).' PVPC across TAs ranges 2.4–75.4%. It is estimated that 600,000–800,000 people are required to achieve economies of scale for viable water-service delivery — a bar small/remote TAs and communities cannot meet alone."
  source_location: "§4.2 Spatial differences, p.44; §4.5 / Discussion, p.46 (600,000–800,000 economies-of-scale threshold)"

NI relevance

This paper was retrieved for RT_316 to answer one question: does the peer-reviewed source actually state the ~NZ23bn drinking-water reticulation network value that [[cr_050_nz-offgrid-infrastructure-cost-reference|CR_050]] cited (and that the ~3,300–5,000/connection was derived from)? It feeds D03.

• **Verification result — 23bn is REAL, but re-framed.** The figure appears verbatim in the Introduction (p.35): *"drinking water reticulation networks could be worth ~NZ23bn”*. CR_050’s attribution is therefore correct — not a mis-attribution. The important caveat for provenance is that this paper repeats a prior estimate rather than measuring it: it derives 23bn top-down as NZ85bn total three-waters value (WICS 2021) × 27% drinking-water-networks share (Water NZ 2022). So the peer-reviewed citation is legitimate, but the number is a soft, derived order-of-magnitude value — and the paper’s own thesis is that it is likely a substantial underestimate (its measured 30.7% poor/very-poor condition is >100% higher than TAs self-reported). Use $23bn as a national top-down floor for the reticulation replacement stock, never as a firm figure.
• The per-connection ($3,300–5,000) is NOT in the paper. RT_316’s framing — “the per-connection is a synthesis derivation” — is confirmed: no per-connection or per-dwelling cost appears anywhere in the source. The paper reports population served (4,135,000) not number of connections, so it does not even supply the denominator for that derivation. The per-connection figure remains an unverified CR_050 back-of-envelope and should not harden a model cell.
• Relation to RT_313 (community reticulation /dwelling or /m) — advanced, not closed. The 23bn national stock confirms the *scale* CR_050 cited but gives no usable per-community reticulation cost. The actual per-metre firming came from the QLDC/Opus Ladies Mile rate card [[ot_113_ladies-mile-dbc-civil-rates-2018|OT_113]] (watermain 200–350/m supply+lay; the CR_050 “1,600–2,500/m watermain" was a mis-attribution of a stormwater rate), with the small-diameter (100–150 mm) community-main rate still open as RT_369. This paper's contribution to RT_313 is contextual: it establishes that NZ's national reticulation stock is large, ageing and under-costed, so a remote community building its own distribution avoids buying into an infrastructure-deficit liability — but it does not itself deliver a /dwelling number.
• The paper’s more directly useful NI content is the condition + small-TA data. The national reticulation stock (57,174 km; 81.1% distribution mains; 67.3% plastic / 19.8% asbestos cement; 30.7% poor/very-poor; mean age 37.8 yr) is asset-replacement context for D03. Most pointedly for NI’s remote-community scope, §4.2 documents the exact bind NI models: small rural TAs with poor pipes and small ratepayer bases (Kawerau 75.4% PVPC on a 7,100 supply population) that cannot reach the 600,000–800,000-person economies-of-scale threshold for viable centralised water delivery — the structural case for the decentralised, community-scale water systems that the D03 evidence base already favours (LIT_007, LIT_065).

Key thesis insights

(Routed context: both — the water-reform / governance content below feeds the thesis argument on institutional capacity and community resilience; the calibration figures feed NI/D03 above.)

• A live NZ case of the governance-scale problem at the heart of remote-community self-sufficiency. New Zealand’s “Local Water Done Well” reform pushes Territorial-Authority water assets into shared Council-Controlled Organisations (CCOs) precisely to reach the estimated 600,000–800,000-person economies-of-scale threshold for viable centralised water-service delivery (§4.5). Small rural TAs — Kawerau (7,100 on public supply), South Wairarapa (7,100), Gore (10,200) — structurally cannot meet that bar, alone or amalgamated with their neighbours. This is the empirical case that decentralised, community-scale water provision is the realistic path for genuinely remote settlements, not a second-best fallback — the governance counterpart to the D03 decentralisation evidence.
• “Invisible infrastructure” and the long-term-governance failure (Boston et al. 2020 framing). Buried reticulation is out of sight, so its renewal is chronically deprioritised under short electoral/budget cycles. The paper’s measured 30.7% poor/very-poor condition is more than 100% higher than TAs self-reported to the regulator Taumata Arowai (13.0%) — Watercare reported 0% while the paper estimated ~40% of its 12,000 km network was PVPC. A governance-attention and accountability failure, not merely an engineering one — material for the thesis chapters on institutional capacity and resilience.
• Infrastructure inequity by deprivation. Condition is worst in high-deprivation areas and small-ratepayer-base TAs, so the communities least able to fund renewal carry the oldest, most-failed pipes — a distributive-justice dimension the thesis’ governance/resilience argument can draw on directly.

Research targets

Documents to retrieve

• (none) — the ultimate primaries behind the 23bn (Water Industry Commission for Scotland 2021; Water NZ 2022 National Performance Review) are deliberately **NOT** raised as new RTs: the 23bn is a soft top-down order-of-magnitude value, and the NI model costs reticulation bottom-up from per-metre rates (OT_113, RT_369). Recorded here as a checked-and-declined provenance note rather than an open target.

Research gaps

• (none new)

Resolved / advanced

• RT_316 (RESOLVED → this page): the tandfonline national assessment is retrieved, read verbatim and ingested; the ~NZ$23bn reticulation network value is verified present (Introduction, p.35) and its provenance clarified (a repeated prior estimate, primaries WICS 2021 + Water NZ 2022, flagged by the paper as likely an underestimate). The per-connection figure is confirmed absent from the paper (a CR_050 synthesis derivation).
• RT_313 (ADVANCED, stays OPEN): national reticulation scale/condition context added; no usable community /dwelling or /m delivered — the per-metre basis remains OT_113 + the open small-diameter residual RT_369.

Notes

Primary peer-reviewed, open-access (CC BY-NC-ND) article — Australasian Journal of Water Resources 29(1):35–48, doi:10.1080/13241583.2025.2481694, published 25 Mar 2025 — read verbatim via pdftotext -layout → data_quality: verified. Every figure in key_claims is quoted verbatim from the raw with page/section locations.

⚠ Load-bearing verification (RT_316) — SUMMARY: the ~NZ23bn drinking-water reticulation network value **IS in the paper** (Introduction, p.35), so CR_050's attribution is **accurate, not a mis-attribution**. But three qualifications matter for any downstream use: (1) it is a **repeated prior estimate**, derived top-down (NZ85bn × 27% ≈ 23bn) from WICS 2021 + Water NZ 2022, not a valuation this paper produced; (2) the paper's central finding is that the deficit is likely a **substantial underestimate**, so 23bn is a floor; (3) the ~$3,300–5,000/connection is NOT in the paper — it is purely CR_050’s synthesis derivation and remains unverified against any primary.

The paper also carries substantial water-reform / governance content (Local Water Done Well, CCO formation, economies of scale, “invisible infrastructure” long-term-governance framing per Boston et al. 2020) that feeds the thesis (community resilience / governance) — captured in ## Key thesis insights above. It is routed context: both: the NI calibration values are the D03 reticulation cost/condition basis, and the governance material feeds the thesis argument, surfaced via the regulation_policy topic.

Connections

Links to

Sources (4): CR_050 · LIT_007 · LIT_065 · OT_113

EDT domains (1): D03: Water, Waste & Circular Systems

Referenced by

Sources (1): CR_050

EDT domains (1): D03: Water, Waste & Circular Systems