Zero Terrain · McWilliams Energy

Underground Pumped Storage — Global Opportunity Screen

Where UPS is buildable (flat terrain, crystalline basement at 500–2,000 m) and needed (fast-growing renewables, storage gap). Adjust the score weights, scan the map, shortlist, then drill into a country.

Screening basis: McWilliams, The Global Need for Underground Pumped Storage Hydro (Aug 2025). Data: Ember 2026 · IRENA 2026-H1 · province-level geology screen.

0

UPS opportunity score

Hover for detail · click a country to open its profile

The two-axis screen

Storage need vs geological suitability · bubble size = projected 2035 VRE · blue = current top 12 by score

Shortlist

Ranked by opportunity score — click headers to sort, rows to open

#CountryRegion ScoreNeedBuild Gap ’35 GWVRE ’35 GW CAGRGeoTerr SeismicConf

Methodology & sources

The combined UPS opportunity score

The score operationalises the screening logic of McWilliams, The Global Need for Underground Pumped Storage Hydro (McWilliams Energy, Aug 2025): a market is attractive where long-duration storage is needed and Zero Terrain-style UPS is buildable. The paper proposes no formal country ranking — this model is our construction from its stated thresholds, and every weight is adjustable in the panel.

Sub-scoreDefinition (0–100)Paper basis
Storage gap 2035log-scaled gap to a storage fleet = 15% of projected 2035 VRE capacity (100 ≈ 300 GW gap)storage need scales with vRE build-out; 15% ≈ today’s leading flexible systems (CN 14%, US 19%, AU 13%)
VRE growth5-yr wind+solar CAGR, capped at 40%/yrvalue rises with vRE growth
VRE shareramp from 5% to 35% of generationstorage need becomes material past ~20–25% vRE penetration
Market scalelog-scaled projected 2035 VRE (100 ≈ 500 GW) market must absorb ≥500 MW / ≥8 h units (~US$2 bn per 1,000 MW / 8,000 MWh)
Geologycurated geology suitability score: crystalline basement reachable at 750–2,000 m, low permeabilitycompetent rock at depth; SBR shafts to 120 MPa UCS, SBC to 250 MPa; groundwater studies gate
Terrain needflatness + absence of conventional PSH resource “zero terrain” — flat land is where UPS’s location-agnosticism has most value

Combination. Need is a weighted mean of its sub-scores. Buildability is a weighted geometric mean of geology and terrain, so flat terrain cannot compensate for missing rock. The final score is the geometric mean of the two axes, so failing either axis fails the screen (the Netherlands has the need but not the rock; Norway the rock but not the need). Penalties: seismicity (high ×0.6, moderate ×0.85), a grid-scale gate (a proxy for absorbing a 500 MW unit, using projected VRE + hydro), and a flexible-hydro discount (reservoir fleets like Brazil’s or Norway’s substitute for storage). Deliverability — market structure, financing, EPC+F/FELT readiness — has no global dataset and is out of scope; read the geology notes and confidence flags alongside.

Renewables, storage & projections

Capacity by technology and country, 2000–2025: Ember Yearly Electricity Data (2026 release; 2024 baseline for 191 of 207 countries). Pumped storage and mixed (pump-back) hydro: IRENA ELECSTAT 2026-H1. Grid batteries: curated 27-market table (NEA/CNESA, EIA, GOV.UK/Modo, Ember Europe, IEA GER 2026; ~96% of global fleet) — other countries treated as ~0 GW and flagged. 2030/2035 projections are decaying-growth extrapolations (r₀ = 5-yr CAGR capped at 25%/yr, decaying 20%/yr): conservative for policy-driven markets — India, Germany and Saudi Arabia project well below stated targets, so their gaps are lower bounds. The storage-gap benchmark (power = 15% of VRE) is a screening heuristic, not a system adequacy study: it ignores duration (GWh), interconnection and demand response.

Geology & terrain screen

Country-level scores curated from geological-province literature (Precambrian shields and cratons, platform basement-depth maps, GSHAP/GEM seismic hazard), terrain ruggedness (Nunn & Puga TRI) and mean elevation, with an adjustment for conventional pumped-hydro resource (ANU RE100 atlas). Country means hide sub-national reality — the “key regions” text in each country profile carries the district-level story (e.g. the USA’s national ~32 hides an excellent Upper Midwest; India’s Deccan pediplains vs the dead Indo-Gangetic plain). Scores are province-level expert judgement, not a basement-depth grid — the confidence flag marks the weakest entries. Water availability and market structure are out of scope (noted qualitatively where decisive).

Siting gates from the paper (what a shortlisted country must still pass)
  • Rock: competent rock at 750 m (Francis variant) or 1,400–2,000 m (deep variant); UCS ≤120 MPa for SBR shaft-sinking, ≤250 MPa with SBC; groundwater studies required.
  • Water: one-off fill of ~550 m³/MWh (−750 m) or ~300 m³/MWh (−1,400 m) — fresh, sea or desalinated; covered upper reservoir suits arid regions.
  • Land: ~1 km² flat or brownfield + 0.25 km² per additional 10,000 MWh.
  • Grid: bi-directional connection ≥500 MW near load centres or congested corridors; economic minimum plant ≥500 MW / ≥8 h (~US$2 bn per 1,000 MW / 8,000 MWh; marginal energy ~US$30/kWh).
  • Timeline: ~3 years studies/licensing + ~5 years construction.