EDT domain: d01_renewable_energy_storage
Definition
Run-of-river micro-hydro converts the energy of a stream (flow × head) into electricity for on-site/community use. Scope here is the community micro/small scale — the model’s micro_hydro technology — not large storage hydro. Output is near-constant when water is available (the model treats it as a flat ~round-the-clock generator, unlike solar/wind), which makes it the most valuable off-grid generator where a viable stream exists. It is physics-gated (D26): most sites simply do not have the flow, head, and proximity to qualify. OT_062
Physics gate & sizing (D26)
- Power:
P = ρ·g·Q·h_net·η, overall efficiency η ≈ 75% (turbine 85% × drive 95% × generator 95% × transformer 98%), net headh_net = h_gross × 0.9. OT_075- Engine note: the NI engine lumps this to a single 0.65 on gross head (≈ 0.75 × 0.9 = 0.675, rounded down — deliberately conservative). Reconciled 2026-07-14.
- Usable flow (environmental limit): the scheme may take only a fraction of the stream — the NES on Ecological Flows sets a residual/environmental flow of 90% of MALF (mean annual low flow) for rivers with mean ≤ 5 m³/s. The model’s canonical gate uses usable flow = 10% of MALF, capping
hydro_kwatP_max = 0.10·MALF × head × 9.81 × η. OT_075- Confirmed by a regulator-facing primary. Cawthron’s advice to Otago Regional Council restates the 2008 proposed NES defaults — minimum flow 90% of MALF + allocation 30% of MALF for rivers with mean flow < 5 m³/s (80%/50% above 5 m³/s) — and proposes tighter Otago defaults (90% minimum flow, 20% allocation, ≤5 m³/s). Essentially every stream the D26 gate admits sits in the <5 m³/s band. OT_100
- The engine’s 10% is conservative, not aggressive: 10% (engine) vs 20% (Cawthron’s proposed Otago allocation) vs 30% (2008 NES allocation) vs >40% = “a high degree of hydrological alteration” (Beca 2008). The 90%-of-MALF minimum flow leaves exactly 10% headroom at MALF, so the engine takes the binding constraint at its cautious reading. OT_100
- ⚠ Ecological risk rises as streams shrink: “the smaller the mean flow, the greater the risk presented by the same flow alteration when defined as a percentage of MALF.” An independent reason for caution on small streams — distinct from the drought-ratio question (RT_334). OT_100
- Third region confirms the 90%-of-MALF default. GWRC’s technical report for its Proposed Natural Resources Plan sets the same default minimum flow — 90% of 7d naturalised MALF at the point of abstraction — adopting the MfE 2008 proposed-NES rule (90% for mean flow ≤5 m³/s, 80% above), identical to Otago’s. Two genuinely new parameters: the return period of MALF is ~1.8 years (a MALF-level low flow recurs roughly every two years, so the environmental-flow cap binds in most years, not rarely); and minimum-flow restrictions may only “become operable” once catchment allocation reaches ≥20% of 7d MALF-N — a de-minimis threshold a community-scale take could sit below. GWRC also declines any extra protection for small streams (90/95/100% of MALF are within ±8% gauging error), a policy divergence from OT_100’s small-stream caution. Consumptive-only, so it does not touch RT_333. OT_160
- Feasibility thresholds (D26): head ≥ 5 m, conveyance distance ≤ 3 km (operational NZ schemes run 0.7–4.5 km, mostly < 3 km). OT_075
- Capacity factor: the model uses
hydro_cf ≈ 0.5(EECA NZ rule-of-thumb; operational small schemes average ~65% — the guide’s figure is conservative). OT_062 OT_075
Drought low flow (D34 stress case)
- Observed NZ drought derate. In the 2019-20 drought the 7-day annual low flow (7dALF) of five large Hawke’s Bay rivers fell to 0.40–0.92 of their long-term 7dMALF — Ngaruroro 0.40, Tukituki 0.47, Mohaka 0.60, Esk 0.68, Wairoa 0.92. The Wairoa catchment escaped the worst of the regional rainfall deficit; excluding it, the drought-affected band is 0.40–0.68. OT_099
- The engine’s
drought_fraction_of_malf= 0.5 (D34 summer-dry gate) sits inside that observed band. It derates the stream’s MALF before the 10%-MALF environmental-flow cap, so a stressed scheme generates from ~5% of MALF. - ⚠ The analogue is large rivers (long-term mean 5.3–77 m³/s), not the community-scale streams the D26 gate admits (feasibility floor 5 L/s). Within that sample, size does not predict the drought ratio — the smallest river (Esk) had the highest ratio (0.68). No small-stream drought statistic is sourced yet (RT_334). OT_099
- ⚠ Abstraction may be banned, not merely reduced. HBRC ceases permission to take surface water (a low-flow ban) once flow falls below the consented minimum flow, and bans were in force in these rivers at the 2019-20 lows. OT_099 A run-of-river scheme is a diversion, and the NPS-FM framework regulates diversions as flow-altering activities alongside dams and abstractions — so it is not exempt merely for being non-consumptive. OT_101 Whether the control is a residual-flow condition or outright cessation is still open (RT_333). If cessation, the D34 summer-dry gate is optimistic for hydro.
Consent load — THREE RMA triggers, not one (REG_012)
The D26 physics gate (flow, head, distance) is a necessary condition, never a sufficient one. A community micro-hydro scheme trips three separate RMA consent triggers, each needing a regional-plan rule or a resource consent: REG_012
| Trigger | What catches the scheme |
|---|---|
| s13 — beds of lakes and rivers | ”No person may… use, erect, reconstruct, place, alter, extend… any structure… in, on, under, or over the bed” or “excavate, drill, tunnel, or otherwise disturb the bed”. An intake or weir is exactly this. |
| s14(2) — water | ”No person may take, use, dam, or divert” water. The diversion itself is caught — expressly. |
| s9 — land use | The penstock and powerhouse are a land use, restricted against NES, regional rule and district rule alike. |
This is a cost and feasibility matter, not paperwork
The wiki previously recorded only the environmental-flow cap. Three consents — bed structure, diversion, land use — is a materially heavier load than the D26 gate implies, and it lands on a technology the model already finds marginal. Any hydro CapEx that ignores the consenting cost is understated.
s14(2)‘s express “divert” also puts RT_333 on statutory footing: a run-of-river diversion is caught by the RMA — confirming from the Act what OT_101 showed from the NPS-FM framework. What remains open is what control applies at low flow (a residual-flow condition, or cessation).
Cost — base $/kW + distance-linear term (D26 / RT_286)
Installed micro-hydro is ~8,000–30,000/kW**, varying substantially with terrain (penstock length, earthworks, consents); the model's `micro_hydro` cost is **19,000/kW (mid of that range). OT_062 D26 splits this into a base machine/intake /kW **+** a **distance-linear penstock term** (`stream_distance × /m`), mirroring grid-extension.
Penstock distance term (RT_286) ≈ ~$90–150/m, medium (trench-dominated):
- Penstock pipe (PE100, AS/NZS 4130): ~$8–13/m incl GST, diameter-dependent — verified NZ retail. URL_018
- Earthworks / trench (dominant): **~
80–140/m** — anchored on the wiki's existing NZ line costs rather than blogs: Electricity Authority build rates35/m overhead /140/m 11 kV underground [[cr_037_nz-rural-grid-connection-costs|CR_037]], and reticulation ~80/m OH / ~300/m UG (incl trenching) [[cr_027_nz-microgrid-reticulation-cost-2026|CR_027]]. Contractor guides (50–120/m basic) corroborate the overhead end. - LV power cable — excluded. The powerhouse→community cable is reticulation (CR_027, ~
8,500/dwelling), already a cost line; a ~8/m hydro cable (URL_019) is documented but not added to the distance term, to avoid double-counting the same wire. - NZTA-accepted bracket (signed primary) — corroborates the live
110/m.** The Ladies Mile Detailed Business Case Estimate (QLDC Housing Infrastructure Fund, June 2018, *accepted by NZTA*) prices bulk excavation **50/m³ + rock extra-over130/m³** (→ a ~**20–65/m open-cut trench floor for a ~0.4–0.5 m³/m penstock trench, inference) and buried welded PE100 pressure main “supply, weld and lay incl. all valves and fittings” at200–360/m** (DN160–DN355) — a trench-floor-to-installed-pipe-ceiling bracket the engine's live `hydro_penstock_capex_per_m` = **110/m (params.py, RT_286) sits comfortably inside, lifting the trench term from corpus-anchored proxies (CR_037/CR_027) to a published NZTA-accepted rate. This updates the ⚠ caution below, which still cites the pre-RT_286 interim $40/m. ⚠ 2018 base, ex-GST, ex-escalation (apply CPI/CGPI); urban-roading context may overstate a rural greenfield trench. OT_113 - The “apply CGPI” escalator is now sourced (RD_032, Stats NZ, verified). The escalation flagged on the OT_113 penstock bracket resolves to the Stats NZ CGPI civil construction series (
CEPQ.S61103, base Sep-2022=1000) = 1101 at Mar-2026, +3.1% YoY; for the trench-dominated term specifically, transport ways (S611031A) = 1094, +2.8% YoY (+2.7% QoQ). Use it to de-date OT_113’s200–360/m installed-pipe /50–130/m³ earthworks rates to the model base year. ⚠ This quarterly release only spans Mar-2024→Mar-2026; the Jun-2018 base for the full 2018→2026 factor must come from Infoshare (CEPQ.S61103) — RT_368 remains open. RD_032
Two modelling cautions (flagged, not auto-applied)
- Engine interim is low.
cost.pyD26 uses40/m** — roughly pipe-only; it omits the earthworks that dominate. A firmer central is **~100–120/m (medium).- Base double-count.
micro_hydro= $19k/kW is the full terrain-dependent installed cost (OT_062 — already embeds a typical penstock/earthworks). Adding a distance term on top counts terrain twice unless the base is redefined as a minimal-penstock (short-run) scheme and the distance term captures the marginal metres. Resolving this split is a model-design decision.
Relevance to Neobiome
Micro-hydro is the NI engine’s micro_hydro technology (D01) — a flat, round-the-clock generator (slice CF ~0.60 across all three slices) gated by the D26 physics thresholds. Where a qualifying stream exists it is the strongest off-grid asset (reaches the evening peak and overnight, cutting storage), which is why the distance-cost term matters: the real cost driver is how far the intake/penstock is from the community, not the machine. Cost inputs: base micro_hydro 19k/kW ([[ot_062_eeca-micro-generation-guide|OT_062]]) + the D26 penstock distance term (RT_286, ~90–150/m medium, trench-dominated). Contributes to I01 (avoided energy cost), I06/I07 (reliable basic-needs energy).
Open questions
- A firm NZ earthworks/trenching rate card (Rawlinsons NZ Construction Handbook, or an EDB/council schedule-of-rates) to lift the dominant trench component from corpus-anchored to a published unit rate (follow-on RT).
- The base-vs-distance split (does $19k/kW already embed the penstock? → redefine base as machine/intake-only) — a model-design decision (D26).
- Per-site MALF and head come from the sampler/GIS; confirm the environmental-flow rule (10% MALF vs NES 90%-MALF residual) is applied consistently.
- What control applies to a run-of-river diversion at low flow? (RT_333 — narrowed) Councils cease consented surface-water takes below the minimum flow (OT_099). A run-of-river scheme is a diversion, and diversions are explicitly regulated flow-altering activities under the NPS-FM framework, alongside dams and abstractions (OT_101, Table 2-3) — so “non-consumptive, therefore exempt” is not available. What remains open is the form of the control: a residual-flow condition on the bypassed reach, or cessation at the minimum flow. If cessation, the D34 stress case should generate nothing below the minimum flow rather than from 10% of the drought flow. Needs a regional-plan-rules answer (HBRC RRMP / ORC RPW / GWRC NRP) or an RMA water-take practitioner.
Connections
Links to
Referenced by
Sources (7): CR_054 · OT_099 · OT_100 · OT_160 · REG_012 · URL_018 · URL_019
EDT domains (1): D01: Renewable Energy & Storage Systems
Concepts (1): Regulatory Consenting for Off-Grid Communities…