OT_099: HBRC State of the Environment 2018-2021 §8 River Flows — observed 2019-20 drought low flows vs 7dMALF

OT_099: HBRC State of the Environment 2018-2021 §8 River Flows — observed 2019-20 drought low flows vs 7dMALF

Source

https://www.hbrc.govt.nz/assets/Document-Library/Reports/SOE-3-Year-report/2018-2021/SOE-2018-2021-Section8-RIVER-FLOWS-digital.pdf — original source (opens in a new tab; the file is not redistributed)

HBRC State of the Environment 2018-2021, §8 River Flows — observed drought low flow

Resolves RT_329 — the observed NZ drought-flow anchor for the D34 summer-dry gate

Section 8 (5 pp) of HBRC’s statutory State of the Environment report. Table 8-1 gives long-term 7dMALF against per-year 7dALF for five monitored rivers, spanning the 2019-20 drought — an observed NZ drought-flow ratio, not a modelled convention. Values read verbatim via pdftotext and recomputed.

Summary

HBRC’s statutory environmental-monitoring report on regional river flows for the hydrological years 2018-19 to 2020-21. Five large rivers (Tukituki, Ngaruroro, Esk, Mohaka, Wairoa) are gauged for annual low flow (7dALF — a 7-day moving average of daily mean flows) against their long-term mean annual low flow (7dMALF). The 2019-20 hydrological year captured a genuine drought (summer/autumn 2020), which makes Table 8-1 a rare NZ record of how far a real river actually falls below MALF in a drought — the constant the D34 summer-dry gate needs. The section also documents the regulatory response: HBRC ceases permission to take surface water once flow drops below a pre-determined minimum flow, and those low-flow bans were in force during the 2019-20 lows.

Key claims

- claim: "Observed 2019-20 drought low flows, as a fraction of long-term 7dMALF (Table 8-1, recomputed): Ngaruroro @ Fernhill 1.68/4.21 = 0.40; Tukituki @ Red Bridge 2.70/5.78 = 0.47; Mohaka @ Raupunga 14.07/23.53 = 0.60; Esk @ Waipunga Bridge 1.45/2.13 = 0.68; Wairoa @ Marumaru 5.41/5.90 = 0.92. Report text (verbatim): 'Annual low flows during 2019-20 were particularly low, ranging from 40% of the mean annual low flow (MALF) for the Ngaruroro River to 92% of mean annual low flow for the Wairoa River. This is likely to be due to the drought that occurred in summer and autumn of 2020.'"
  source_location: "Table 8-1 (p.69) + body text p.69"
- claim: "The Wairoa outlier (0.92) is climatic, not hydrological: 'the Wairoa catchment was less severely impacted by lower rainfalls than the rest of the region, hence the smaller impact on the mean annual low flows.' Excluding Wairoa, the drought-affected band across the other four rivers is 0.40-0.68."
  source_location: "p.69"
- claim: "Climate, not abstraction, drives the Mohaka/Wairoa shortfall: 'Relatively small allocations of surface water are consented for abstraction from these two rivers (e.g., Wairoa 0.32 m3/s), so the difference between the mean annual low flow ... and the annual low flow for the 2020 summer/autumn ... mainly reflects the climate impact on these river flows, rather than anthropogenic influences.'"
  source_location: "p.69"
- claim: "Low-flow bans cease consented takes at the minimum flow: 'HBRC manages the effects of surface water takes on low flows in the Tukituki, Ngaruroro, and Esk Rivers (which relative to the Mohaka and Wairoa rivers have a much higher number of takes) by ceasing permission to extract water (low flow ban) when river flows are less than a pre-determined threshold, called a minimum flow.' And: 'Figure 8-3 shows that low flow bans were in place when the lowest flows occurred in these rivers during 2019-20.'"
  source_location: "p.70-71 (Figure 8-3)"
- claim: "Definitions: ALF is the annual low flow, calculated from a 7-day moving average of daily mean flows for each hydrological year; MALF is the mean (average) of ALF statistics from all years of flow record. Hydrological years run July to end of June, chosen 'to avoid splitting low flow periods in the statistical analyses.'"
  source_location: "footnotes 1-3, p.68-69"

Neobiome Intelligence relevance

Resolves RT_329. Upgrades drought_fraction_of_malf (engine params.py, the D34 summer-dry gate) from assumed to sourced: the model’s 0.5 now sits inside an observed NZ drought band (0.40-0.68, excluding the rainfall-spared Wairoa) rather than being a bare engineering guess. In the D34 stress pass this fraction derates stream_flow_l_s (which the model treats as MALF) before the 10%-MALF environmental-flow cap, so it bites both gates — the winter gate (hydro under-delivering its winter commitment) and the summer-dry gate (the dry-year water balance). Feeds micro_hydro and d03_water_waste_circular (where it bounds the currently-inert stream_abstraction lever, limitations register O2).

Two limits on transferability — both material

1. The analogue is large rivers. The five gauged rivers have long-term mean flows of 5.3-77 m³/s; the model gates community-scale streams (D26 feasibility floor 5 L/s). Nothing here evidences how a small headwater stream behaves in drought, and within this sample, size does not predict the ratio — the smallest river (Esk, mean 5.28 m³/s) had the highest drought ratio of the three abstraction-managed rivers (0.68), while the lowest ratio (0.40, Ngaruroro) came from a large one. Do not read 0.40-0.68 as “smaller stream → lower fraction”. → RT_334.
2. One drought year, not a return period. 2019-20 is a single observed event, not a 1-in-N statistic. Pearson (1995) Regional Frequency Analysis of Low Flows in New Zealand Rivers, J. Hydrology (NZ) 33(2):94-122 remains the canonical NZ low-flow frequency reference (paywalled, JSTOR).

Open modelling risk — abstraction may be banned, not merely reduced

HBRC ceases consented surface-water takes below the minimum flow, and the bans were in force at the 2019-20 lows. The engine’s stress case instead lets micro-hydro generate from 10% of the drought flow (≈5% of MALF). A consumptive take would be at zero. Micro-hydro is non-consumptive (run-of-river, water returned below the tailrace) so it plausibly escapes a low-flow ban — but that is an untested assumption, and if it is wrong the D34 summer-dry gate is optimistic for hydro. → RT_333. Recorded, not acted on.

Research targets

Documents to retrieve

• RT_335 — GWRC Minimum flow recommendations for the Wellington region (technical report supporting the Proposed Natural Resources Plan): Q5 (1-in-5-yr 7-day low flow), 90%-Q5 and 60%-of-1-in-20-yr conventions, MALF return period ~1.8 yr.

Research gaps

• RT_333 — Does a non-consumptive run-of-river hydro take face the low-flow ban that a consumptive take does? Determines whether the D34 stress case may generate at all below the minimum flow.
• RT_334 — A NZ low-flow statistic for small / headwater catchments (community scale), to replace the large-river analogue this source provides. Candidates: NIWA low-flow statistics (Singh et al. 2021), Pearson (1995) frequency curves.

Notes

Authoritative HBRC publication (not AI-prepared) — data_quality: high. Retrieved 2026-07-14. Table 8-1 was re-verified against the PDF (pdftotext -layout) and all five ratios recomputed independently.

Framing the band as “rain-fed 0.40-0.68 / high-baseflow 0.92” is not this source’s language — HBRC attributes Wairoa’s 0.92 to that catchment escaping the regional rainfall deficit, not to baseflow character. The high-baseflow-vs-rain-fed distinction belongs to the Cawthron/ORC advice memo (OT_100), which defines it explicitly.

Connections

Links to

Sources (1): OT_100

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

Technologies (1): Micro-hydro (run-of-river, community-scale)

Referenced by

Sources (4): CR_053 · OT_100 · OT_101 · OT_160

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

Technologies (1): Micro-hydro (run-of-river, community-scale)