H1 2026 Evaluation
New Zealand Reservoir Transparency
C+60Below Average — Ranked #42 out of 167 countries
weight 30%
weight 20%
weight 15%
weight 13%
weight 10%
weight 8%
weight 5%
Primary source evaluated
Electricity Authority — Hydrological Modelling Dataset (EMI / Data & Insights Hub)
https://www.emi.ea.govt.nz/Environment/Datasets/HydrologicalModellingDatasetDimension breakdown
Data Availability
20% of total score
For the COVERED subset (~22 reservoirs: HMD's 6 major lakes + Meridian's 7 + Mercury's 9 Waikato lakes, with overlap), hydro storage data is publicly available through multiple channels. The Electricity Authority's HMD is freely downloadable covering storage, spill and flows from 1932 to 2024. Aggregate daily storage (NZ and South Island) downloadable from EMI historical risk curves report. Meridian Energy publishes weekly PDF lake levels for 7 lakes; Mercury publishes 30-minute levels for 9 Waikato lakes. Near-real-time individual lake storage data from NZX Hydro requires paid subscription. The typical covered reservoir has weekly to daily public visibility but with paywalls for the most timely individual-lake data.
Technical Accessibility
15% of total score
For the COVERED subset, the dominant access mechanism is bulk file downloads (HMD CSV/Excel) plus scrapable HTML/PDF. The Electricity Authority documents two REST APIs (ICP connection data and real-time wholesale dispatch prices) via Azure developer portal, but neither provides hydro storage. ECAN operates a developer API portal requiring registration and a subscription key. NIWA's Hydro Web Portal offers an Export API URL feature but is not a fully documented public REST API for lake storage. No country-level open hydro storage REST API exists.
Coverage
30% of total score
Conservative capacity-weighted estimate applied 2026-05-29. Numerator/denominator: n_covered ≈ 20 / n_total ≈ 50 reservoirs >10 hm³ per GRanD and the NZ Large Dam Register. Genesis, Mercury and Meridian publish data covering the major hydropower lakes (HMD's 6 major + Meridian's 7 + Mercury's 9 Waikato cascade, with overlap = ~20–22 unique). Smaller council-managed reservoirs (Wellington's Wainuiomata, Christchurch supply dams, Auckland Hunua and Waitākere ranges) are tracked individually by regional councils via LAWA but not aggregated nationally with storage data. Applying the conservative downward adjustment (~-4 points) to recognise the structural gap of smaller council/supply reservoirs not captured in the hydropower-driven publication framework: coverage = 40. The covered subset still represents the dominant hydropower storage but the long tail of supply reservoirs above the 10 hm³ threshold remains structurally invisible.
Historical Depth
13% of total score
MEDIAN historical depth across the ~22 covered reservoirs is approximately 20+ years. The HMD provides storage and flow data back to 1932 (~92 years) for the 6 major generation lakes — exceptionally deep. NZX Hydro additionally holds historical data since 1980 for 14 catchments. Meridian and Mercury portals expose shorter histories (~10-15 years) for their respective lakes. With 6 reservoirs at 90+ years, 14 at 40+ years via NZX, and the Waikato cascade with 10-15 years, the median lands clearly in the 20+ years band. Band: median 20+ years.
Update Frequency
10% of total score
For the COVERED subset, the typical free-tier cadence is WEEKLY. The EMI/EA historical risk curves report updates aggregate hydro storage daily. Meridian's comparative lake levels PDF is updated weekly. Mercury's Waikato lake levels update every 30 minutes (web display only). Transpower's hydro charts update weekly. The HMD bulk dataset updates annually or semi-annually. Near-real-time individual lake storage is paywalled via NZX Hydro. The free tier offers a mix of daily aggregate and weekly individual lake PDFs.
Methodological Transparency
8% of total score
For the COVERED subset, the Electricity Authority publishes explanatory articles documenting storage in GWh (energy equivalent), distinguishing 'controlled storage' (always usable) from 'contingent storage' (available under shortage declarations), linked to resource consent operating levels. Capacity figures (Pukaki 1,595 GWh, Tekapo 770 GWh, national total ~4,500 GWh) are publicly cited. The HMD documentation explains dataset structure. However, no single consolidated open methodology document details how raw lake level measurements are converted to GWh, how capacity curves are derived, or how consent conditions translate to storage bounds.
Language and Usability
5% of total score
All data portals, reports, APIs, documentation, and publications are exclusively in English. New Zealand is a monolingual English-speaking government environment for all official data infrastructure. No translation barriers exist for any of the described sources.
Evaluator notes
New Zealand has one of the world's strongest frameworks for publishing hydroelectric storage data, driven almost entirely by electricity market necessity rather than water resource management mandates. The Electricity Authority's HMD is exceptional by global standards: freely downloadable, covering major generation lakes from 1932 at daily resolution, with no registration required. Under methodology v1.2.0 strict linear coverage, NZ scores only 44 because the national reservoir inventory (~50 dams >10 hm³ per GRanD) is much larger than the ~22 hydro generation reservoirs that have public structured data — many regional water-supply dams above the threshold are not in any national publication. Historical depth scored on the MEDIAN covered reservoir lands in the 20+ year band thanks to HMD's exceptional 92-year archive for the 6 major lakes. The critical gap is the bifurcation between free historical data and paywalled near-real-time data: the authoritative operational lake storage series from NZX Hydro (NIWA data) requires a paid subscription. NZ would rank substantially higher if NZX Hydro data were made freely accessible or if the EA added a hydro storage endpoint to its existing APIs and if regional water-supply reservoirs were rolled into a national publication.
Evaluated by Jaime Delgado · 2026-05-29 · Methodology v1.3.0