Hydropower: From Water Cycles to Grid-Scale Energy and Storage

9.1 — Hydropower is solar energy, once removed

The Sun evaporates water from oceans and land. Vapour rises, cools, condenses and falls as rain and snow — some of it onto ground that is a long way above sea level. Gravity then does the rest, and a turbine intercepts the water on its way down.

So hydropower is solar energy that has already been collected, concentrated and lifted for free by the atmosphere. The collector is a river catchment measured in thousands of square kilometres; the concentration ratio is enormous; and unlike a photovoltaic panel, the collected energy arrives already stored.

Unlike coal, gas or uranium, the primary resource is not consumed. The water passes through the machine and continues downstream, and the atmosphere lifts it again. This is why hydro plants routinely operate for eighty or a hundred years: there is no fuel cost to erode the economics and no combustion to wear out the hot section.

9.2 — The governing equations

Two variables set everything: flow, the volume of water passing per second, and head, the effective vertical drop. They multiply, so a given power output can be reached by many combinations of the two — and that is precisely where the engineering difficulty lies.

Note also what the equation does not contain: time. Chapter 3’s distinction between power and energy applies with unusual force here. P = ρgQHη tells you the plant’s power. How long it can sustain that power is a question about the reservoir, not the turbine — and those two numbers are set by completely different parts of the project.

9.3 — The basic plant architecture

Water leaves the reservoir through an intake, travels down a penstock under increasing pressure, gives up its energy in the turbine, and exits through the tailrace back into the river. The turbine shaft turns a synchronous generator; a step-up transformer raises the voltage for transmission, exactly as in the thermal plants of Chapter 1.

The comparison with Chapter 8 is instructive. A nuclear station is a conventional steam plant with an exotic heat source. A hydro station dispenses with the thermodynamics altogether: there is no boiler, no steam cycle, no condenser, no Carnot limit. Water pushes a runner directly. This is why hydro efficiencies of 90% and above are ordinary, while a nuclear plant struggles past 33% — hydro is not a heat engine, so the penalty that applies to every heat engine does not apply to it.

9.4 — Turbine selection: matching machine to site

9.5 — Reservoir hydro versus run-of-river

The distinction matters more than the label suggests. Run-of-river hydro is a variable renewable: predictable across seasons, uncontrollable within a day. Reservoir hydro is firm, dispatchable capacity. Two plants on the same river with the same nameplate rating can occupy entirely different positions in a grid operator’s merit order, and be worth entirely different amounts, purely because one of them has a wall.

9.6 — The reservoir as a natural battery

There is one further property worth noticing: the storage does not leak. A lithium-ion cell self-discharges. A reservoir loses water to evaporation and seepage, but the loss is slow enough that seasonal storage is genuinely practical — some schemes fill with snowmelt in spring and discharge through the following winter. No electrochemical technology in commercial use comes close to holding energy for six months.

9.7 — Capacity factor

A 1,000 MW hydro plant does not generate 1,000 MW × 8,760 hours a year. Actual output depends on rainfall, reservoir management, environmental release requirements, irrigation and drinking-water obligations, and commercial strategy. Reservoir hydro typically lands in the 35–55% range.

9.8 — Peaking power, ancillary services and black start

Hydro turbines can go from standstill to full load in a couple of minutes and can change output within seconds once running. That makes them exceptionally valuable for services that have nothing to do with bulk energy: frequency regulation, spinning reserve, load following, voltage support, and the synchronous inertia discussed in Chapters 2 and 5.

One capability deserves separate mention. A hydro plant can perform a black start — restarting with no electricity from the grid at all. Open the gates, and the water does the work. Nearly every other generator needs external power to begin: a thermal plant needs pumps and fans, a nuclear plant needs its safety systems energised, a solar farm needs its inverters to see a grid to synchronise with. After a system-wide collapse, the sequence that rebuilds a national grid usually starts at a dam.

9.9 — Pumped storage: electricity to water to electricity

Pumped storage uses two reservoirs at different elevations. When electricity is cheap, the plant consumes it to pump water from the lower reservoir to the upper one. When electricity is expensive, it releases the water back down through turbines. Modern machines are usually reversible pump-turbines: the same runner does both jobs, spinning one way to generate and the other to pump.

This is the arbitrage logic of Chapter 4, implemented in gravity rather than electrochemistry. And it is not a niche: pumped storage remains the overwhelming majority of the world’s installed grid-scale storage by energy capacity, despite two decades of battery deployment.

9.10 — Round-trip efficiency and the price ratio

Nothing about the cycle is free. Pumping losses, friction in the waterways, turbine losses and generator losses compound, putting round-trip efficiency in the 70–85% range. Roughly a fifth of the electricity that goes in never comes back out.

9.11 — Pumped hydro against batteries

The last row of that table is the one that decides the long-run division of labour. Extending a battery from four hours to eight means buying twice the cells: energy scales with cost almost linearly. Extending a pumped-storage scheme from seven hours to fourteen may mean raising a dam wall a few metres, because reservoir volume grows far faster than the structure containing it. Batteries are therefore the cheap way to buy power and the expensive way to buy energy; pumped hydro is the reverse.

They are complements, not competitors. A battery handles the sub-second frequency response and the sharp two-hour evening spike. A pumped-storage scheme handles the multi-day weather event that no economically sized battery fleet can cover.

9.12 — Dams as multi-purpose infrastructure

Most large dams serve several functions at once: water storage, flood control, irrigation, municipal drinking water, navigation, and electricity generation. This makes a hydro project multi-purpose infrastructure that happens to generate power, rather than a power plant that happens to hold water.

9.13 — The civil engineering side

Every dam also needs a spillway: a route for floodwater to pass safely around the turbines during extreme inflow. It is worth being clear about why. An embankment dam that is overtopped can erode and fail within hours. The spillway is not an efficiency feature or an operational convenience; it is the structure that keeps the dam from destroying itself, and it must be sized for a flood far larger than any in the historical record.

9.14 — Sedimentation

Rivers carry sand and silt. A dam slows the water, the sediment settles, and the reservoir gradually fills with rock. This is the one degradation mechanism unique to hydro, and it is the reason a hydro plant’s energy capacity and its power capacity age at completely different rates.

Management options exist — drawdown flushing, sluicing during high flow, dredging, upstream catchment measures — but all are expensive, most require sacrificing generation, and some export the problem downstream. With much of the world’s hydro fleet now past mid-life, sediment is shifting from a footnote in feasibility studies to a live question about how much storage a country actually still has.

9.15 — Environmental impact and resettlement

Large dams alter river systems fundamentally. They block fish migration, trap the sediment that would otherwise replenish downstream floodplains and deltas, change the temperature and oxygen content of released water, shift flow patterns that ecosystems evolved around, and inundate land that people live on and farm.

Resettlement is the hardest of these. Filling a reservoir can displace communities whose livelihoods, social structures and land rights are bound to a specific valley. Compensation and livelihood restoration are genuinely difficult to do well, and the record across many large projects is poor.

9.16 — Geological risk and climate change

A dam transfers enormous forces into the surrounding rock. Foundation quality, fault location, abutment strength and seismicity all have to be characterised before the design can be fixed, and site investigation is one of the few line items where underspending reliably costs more than it saves.

The practical result is that hydrological assumptions have become a live financing risk rather than a settled input. A project underwritten on a hundred-year flow record is being underwritten on a climate that no longer exists.

9.17 — Hydro as a financial asset

Water in a reservoir behaves like inventory. Every hour, the operator faces the same question: generate now at today’s price, or hold and generate later at a price that might be higher. Because the water can be held at almost no cost and released at will, the decision has the structure of an option rather than a production schedule.

9.18 — Pumped hydro project economics

Pumped storage shares nuclear’s cost structure more than it shares a battery’s: very high capital cost, very long construction, negligible fuel cost, and a payback period measured in decades. Chapter 8’s lesson therefore transfers directly — the cost of capital and the construction schedule dominate the economics, and a project delayed is a project made more expensive whether or not anything went wrong technically.

The revenue side is the harder problem. Arbitrage income depends on price spreads that nobody can forecast twenty years out, and a scheme financed on today’s spreads is exposed to the possibility that cheap batteries compress exactly the spread it was built to capture. This is why capacity mechanisms and long-term contracts matter more to pumped storage than to almost any other asset: without them, a project with a fifty-year life is being asked to finance itself on a five-year view of the market.

9.19 — Power capacity against energy capacity

Chapter 3 introduced this distinction. It is worth restating here because hydro is where it bites hardest: the power rating is set by the turbines and the penstock, the energy rating by the reservoir, and the two are built by different disciplines on different budgets. A quoted “1 GW pumped-storage project” is not a specification. It is half of one.

9.20 — Hydro and transmission geography

Large hydro resources sit where the mountains and the rain are, which is rarely where the demand is. A Himalayan or Andean scheme may be several hundred kilometres from the nearest industrial load, and the transmission line to reach it can rival the plant in cost and exceed it in permitting difficulty.

Chapters 6 and 7 made the same point about solar and wind, and the pattern is now general enough to state plainly: in a decarbonising grid, the binding constraint on new generation is increasingly the wire rather than the resource. The best sites were never the closest ones, and the ones that remain are further still.

9.21 — India’s hydro opportunity

India has substantial hydropower potential in the Himalayan river systems, the Western Ghats and the north-east, much of it undeveloped. The strategic case is not primarily about energy volume — it is about shape.

The honest summary is that hydro solves a problem India genuinely has, in a way batteries alone cannot afford to at multi-day scale — and that the obstacles are almost entirely outside the engineering.

9.22 — The deeper lesson

Hydropower demonstrates something the rest of this series has treated as an assumption: generation and storage do not have to be separate assets. One reservoir is simultaneously a power plant, an energy store, a water supply, a flood-control structure and an irrigation source. That multifunctionality is what makes it strategically important, and it is also what makes it politically difficult.

• •Physics — head and flow multiply to give power, but they are not interchangeable in cost. Head is bought from geography; flow is bought from concrete.
• •Machine — Pelton, Francis or Kaplan follows from the site, not from preference.
• •Architecture — reservoir, run-of-river or pumped storage. This decides whether the asset is dispatchable at all.
• •Duration — energy divided by power. The most commonly omitted number in any hydro announcement.
• •Efficiency — for pumped storage, round-trip losses set a required price ratio, not a required spread.
• •Sediment — the mechanism that erodes energy capacity while leaving power capacity untouched.
• •Competing uses — irrigation, drinking water, flood control and environmental flows frequently outrank electricity.
• •Consent — resettlement and ecological impact are the most common reason good projects do not get built.
• •Hydrology — the historical flow record is no longer a reliable forecast of the future one.
• •Transmission — the resource is where the mountains are; the demand is not.

Chapter 10 turns to the generation technology that still supplies more of the world’s electricity than any other, and against which every source in this part has implicitly been measured: combustion. Coal and gas — the incumbent, its economics, and why the cheapest plant to build is so often the most expensive to run.