Run-of-River vs Reservoir vs Peaking: Hydropower Types in Nepal

Nearly every hydropower plant in Nepal's database carries a one-word classification: run-of-river, peaking run-of-river, reservoir (storage), or pumped storage. This 'type' is not a technical footnote. It is the single most important label for understanding when a plant can actually deliver electricity, how much it earns per unit, and how it behaves across Nepal's sharply seasonal year. A 100-megawatt (MW) run-of-river plant and a 100 MW reservoir plant have the same nameplate rating but very different value to the grid.

The distinction comes down to storage. All the types capture the energy of falling water passing through turbines, but they differ in how much water they can hold back and for how long. That storage window, from a few hours to several months, decides whether a plant is a slave to the day's river flow or a manager of it. The Department of Electricity Development (DoED) and the Nepal Electricity Authority (NEA) use these categories in licensing, tariff-setting and power purchase agreements (PPAs).

Because Nepal's rivers are fed by monsoon rain and Himalayan snowmelt, flows swing enormously between the wet monsoon (roughly mid-June to mid-September) and the dry winter and pre-monsoon months. Understanding plant types is therefore the key to Nepal's most stubborn energy puzzle: how a country that exports surplus electricity every monsoon still imports power from India every winter.

A run-of-river plant has little or no water storage. It diverts a portion of the river's natural flow through a headrace and turbines, then returns it downstream. Whatever the river carries at that moment sets the output: when the river is high, the plant runs near its rated capacity; when the river shrinks, generation falls with it. RoR plants use a small diversion weir rather than a large impounding dam, so they flood less land and cost less to build, which is a major reason they dominate Nepal's fleet.

The trade-off is seasonality. During the monsoon, most RoR plants run at or near full capacity and Nepal produces a surplus. In the dry months, river discharge in mid-February to mid-March can fall to roughly 30 to 40 percent of the annual average, and many pure RoR plants generate only 20 to 40 percent of their nameplate rating. This is why RoR capacity figures overstate what the fleet reliably delivers in winter, and why the type field is so important when reading any plant's numbers.

Run-of-river is by far the most common type in Nepal, accounting for roughly 90 to 98 percent of installed hydropower depending on how partially-regulated plants are counted. Well-known examples include the 456 MW Upper Tamakoshi (Nepal's largest completed plant, with limited peaking pondage), the 144 MW Kali Gandaki 'A', and hundreds of small and medium private plants built by independent power producers (IPPs).

A peaking run-of-river plant, written PRoR or PROR, is a run-of-river plant fitted with a small storage basin called a pondage. The pondage is too small to store water across seasons, but it can hold a day's (or sometimes a week's) inflow. The plant collects water during off-peak hours, then releases it through the turbines during the few hours of highest demand, typically the evening peak. A common engineering rule of thumb is that a pondage plant can run at full output for up to about 24 hours' worth of stored water before it must rely on live inflow again.

PRoR matters most in the dry season. When a river's natural flow is below a plant's design discharge, a pondage lets the operator concentrate that limited water into a few hours of firm, dispatchable power rather than spreading it thinly across the whole day. Because this evening peaking capacity is more valuable to the grid, NEA pays PRoR (and reservoir) projects a higher tariff than plain RoR plants; historically a plant able to run several hours at full capacity in the dry season has earned a premium dry-season rate.

Several major Nepali plants are peaking run-of-river. The 144 MW Kali Gandaki 'A', NEA's largest plant for years, is designed with about six hours of daily peaking pondage. The 900 MW Arun-3 (under construction) is conceptualised as a PRoR scheme. Peaking pondage is a middle path: cheaper and less disruptive than a full reservoir, but with enough flexibility to firm up the evening peak.

A reservoir or storage plant impounds a large dam and holds water for weeks or months, not hours. It captures abundant monsoon runoff behind the dam and releases it through the dry season, so the plant can generate firm, on-demand power when RoR plants are starving. This seasonal carry-over is exactly the capability Nepal's grid lacks, and it is why reservoir projects are treated as a national priority despite their higher cost, longer construction time and the land and communities they displace.

Kulekhani is Nepal's only true reservoir (storage) hydropower scheme. Its Kulekhani Dam, about 114 metres high, created the Indra Sarobar reservoir, and Kulekhani-1 (60 MW) was commissioned on 8 December 1982 as the country's first storage plant. Water released from the reservoir cascades through Kulekhani-1, Kulekhani-2 (32 MW) and later Kulekhani-3 (14 MW), giving a combined installed capacity of about 106 MW. The reservoir's design storage was roughly 85 million cubic metres, though sediment, including debris washed in by past disasters, has reduced usable storage over the decades.

Because Kulekhani stores monsoon water, it is uniquely able to run at full output during winter evening peaks when almost every other Nepali plant is throttled by low flows. That is precisely why one storage scheme is not enough for a grid of over 3,400 MW: a single 106 MW reservoir cannot cover the winter deficit of a fleet that is otherwise almost entirely run-of-river.

Pumped storage hydropower (PSH) is a distinct fourth type. It uses two reservoirs at different elevations. During periods of surplus, cheap electricity (in Nepal, the monsoon daytime glut), the plant pumps water from the lower pond up to the upper pond, consuming power. During peak demand, it lets that water fall back through the turbines to generate. It is a net consumer of energy over a cycle but acts as a large grid-scale battery, storing surplus generation for when it is needed.

Nepal has no operating pumped storage plant, but studies, including work highlighting the Himalaya's steep terrain and existing reservoirs, point to strong potential. As solar generation grows and monsoon hydropower produces mid-day surpluses that currently have to be exported cheaply or spilled, pumped storage is increasingly discussed as a way to time-shift that energy into the evening peak. For now, treat any PSH label in the database as a proposed or study-stage project rather than an operating plant.

The seasonal paradox follows directly from the type mix. With roughly 90 to 98 percent of capacity being run-of-river, national output rises and falls with the rivers. In the monsoon, RoR plants run near full rating, domestic demand is comfortably met, and NEA exports the surplus to India and, since 2025, to Bangladesh. In fiscal year 2024/25, Nepal exported around 2,380 gigawatt-hours (GWh) of electricity, the second year it exported more than it bought back.

In winter and pre-monsoon, the same RoR fleet collapses to a fraction of its capacity just as heating and lighting demand climbs toward the evening peak. To keep the lights on, NEA imports power from India: in FY 2024/25 it imported roughly 1,700 GWh, worth about Rs 12.9 billion, concentrated in the dry months. The country is a net exporter over the year but a structural importer in the dry season, and industrial consumers can still face load management. This is the concrete answer to the common search 'why does Nepal have dry-season power cuts': too much run-of-river, too little storage.

The fix is not more nameplate megawatts but more storage megawatts. Reservoir and peaking-reservoir projects convert monsoon surplus into firm winter and peak energy, reducing both dry-season imports and monsoon spillage. That is why the type field, not just the MW rating, is the honest measure of what a plant contributes to Nepal's energy security.

Two large storage projects would transform the picture. The proposed Budhi Gandaki Hydroelectric Project, a 1,200 MW reservoir scheme on the Budhi Gandaki river roughly 80 km by road from Kathmandu, is planned around a very high dam (about 263 metres) creating a reservoir tens of kilometres long. Its steady, regulated release would let it deliver firm power year-round and generate on the order of 3.3 to 3.4 billion units (kWh) annually. As a single project it would dwarf Kulekhani's storage many times over.

The 635 MW Dudhkoshi Storage Hydroelectric Project in Koshi Province (Khotang, Okhaldhunga and Solukhumbu districts) is a seasonal-reservoir scheme backed by proposed Asian Development Bank (ADB) financing. It pairs a large main powerhouse with a smaller ecological-flow station and is expected to generate roughly 3,400 GWh a year, with commercial operation targeted around the early 2030s. Together with Budhi Gandaki, it would give the grid the dry-season and peaking backbone that a single Kulekhani cannot provide.

The practical effect of adding this storage would be to flatten Nepal's seasonal see-saw: less winter import from India, less monsoon energy spilled or dumped for export at low prices, and a firmer platform for integrating variable solar. In short, storage plants are what let a monsoon-rich, dry-season-poor grid finally match supply to demand across the whole year, which is why they are named as national priorities even though they are slower and costlier to build than the run-of-river plants that make up almost all of today's fleet.