Works in Progress is becoming a print magazine. Our first print issue, Issue 21, will land in November. If you live in the United States or the United Kingdom, you can subscribe here. If you live outside the US or UK and want to be notified as soon as subscriptions are live in your country, leave your details here.
When India suspended the Indus Waters Treaty in April 2025, it ended the world’s most successful water-sharing accord. The 1960 treaty had survived wars and coups, guaranteeing Pakistan unhindered flows from three of the six rivers that make up the Indus system. Even limited upstream impoundment could choke the water that irrigates four fifths of Pakistani farmland. Pakistan’s Prime Minister called the action an act of war.
But without new infrastructure, India can only partially make good on its threats. To give itself the capability to desiccate the river’s dependents at will, it needs more and larger dams.
Many other nations are actively arming themselves with the infrastructure to make these types of threats. Dams pose an increasingly pressing threat to people who lie downstream of them all over the world. The issue is clearest in Tibet, where China controls chokepoints that control the freshwater resources that some two billion people in downstream nations rely on for transportation, drinking water, irrigation, and industry.
There is a technological solution for these people. Cheap, accessible desalination can reduce water scarcity for both upstream and downstream nations without them having to depend on diplomacy or worse.
Why we dam
Dams are typically built for three main reasons. First, to store water for later. By storing water behind a dam instead of letting it flow by, operators can better match supply to demand.
Second, to control its flow and generate power. For hydroelectric dams, the impounded water is a means to store potential energy. The height difference between the water level behind and in front of the barrier determines how much energy it can create. The power of the rushing water can be used to spin a turbine, creating hydroelectricity.
Hydroelectric generation is especially valuable because it creates little carbon dioxide, and because it is highly flexible – the floodgates or valves that feed the turbines can be opened or closed at will to respond to demand smoothly. When not in use, the energy is easily stored for later as potential of the water behind the dam. In this configuration, the dam and its turbines act as a huge battery. Stored hydro is the only battery type that has worked at a large scale so far.
About 15 percent of the world’s total electricity – approximately 4,200 terawatt hours per year – is generated this way. China is the world’s largest hydroelectric generator, generating 1,226 terawatt hours per year, or just over 29 percent of global hydroelectricity.
Third, dams can also divert water rather than storing it or controlling its flow volume. The dams of the Catskill Watershed divert water through massive tunnels deep underground, running underneath the Hudson River and into the Croton Watershed, where the water is collected and treated before feeding into New York City’s pipe network. They provide a combined one billion gallons per day, or about 75 percent of New York City’s drinking water. Without this infrastructure, the water would instead flow into the Delaware River, meeting the sea 120 miles further down the coast.
Dams are powerful pieces of infrastructure with many benefits. However, when a dam is placed on an international river, there is no governing body tasked with ensuring the benefits accrue to upstream and downstream populations equally. The dam builder has near-total control. Downstream countries, in turn, have little recourse to right this imbalance. Bringing countries with headwater control to the negotiating table is difficult when they have all the leverage. As a result, countries reliant on the upstream nation are often forced to make concessions.
Consider the Grand Ethiopian Renaissance Dam, built between 2011 and 2023, one of the biggest hydroelectric infrastructure projects in Africa. It has the largest reservoir volume of any transboundary dam, at up to 74 billion cubic meters of water. Constructed on the Blue Nile river, the dam is designed to provide 5.15 gigawatts of electricity to the Ethiopian grid.
While the dam sits upstream of much of the Blue Nile’s catchment basin, if the Ethiopians started with an empty reservoir and began filling it to its full capacity they could still reduce the total flow of water in the Nile by 39 percent for up to 18 months.
The large volume of water contained in the dam cushions Ethiopia from drought and allows its hydroelectric turbines to run at a consistent rate, providing a steady backbone to the Ethiopian power grid. But the dam also allows Addis Ababa to manipulate when water is released downstream. If released suddenly, it could inundate the low-lying farms and cities of Sudan and Egypt and overflow many of the older and shorter dams downstream. This would leave the reservoir empty and ready to be filled during the next dry season, just when those same farms and cities downstream need water the most.
The threat of this one-two punch has caused Egypt’s attempts at stopping the Grand Ethiopian Renaissance Dam’s filling process to become increasingly aggressive. The Ethiopian government has installed an umbrella of Russian and Israeli air defense missile platforms to protect the dam from a possible airstrike. As of September 2024, the dam has been filled, and Ethiopia has gained a bargaining chip big enough to reshape the regional balance of power.
Egypt claims that the Grand Ethiopian Renaissance Dam is an attempt by the Ethiopian government to gain the diplomatic upper hand against downstream nations. Independent engineering commissions have found that the dam and its reservoir are oversized compared to the average flow of the Blue Nile and the flow that would be required to power the dam’s turbines adequately.
There are many other examples of transboundary rivers whose management is a point of conflict between countries. The Daryan Dam diverts up to 60 percent of the water from the Sirvan River – a tributary of the Tigris that contributes 20 to 30 percent of the river’s total flow – from Iraq to southwestern Iran via a 48-kilometer tunnel. Iraq, which relies heavily on the Tigris and Euphrates Rivers, opposed the construction of the Daryan Dam but was unable to exact concessions.
He who controls the Tibetan plateau
Nowhere is the impact of damming shared rivers more pronounced than in Asia. An estimated 718 billion cubic meters of water flows from the Tibetan Plateau every year. Of this, only 35 percent – roughly 250 billion cubic meters per year – flows wholly through China. 48 percent of Tibet’s water supply flows into India through rivers like the Brahmaputra and Ganges and, to a lesser degree, into Pakistan through the Indus. Whichever nation controls the Tibetan Plateau controls the water supply of nearly two billion people.
Controlling the flow of that water is uniquely easy. Glaciers on the plateau accrue water during the region’s winters, when water demand is at its lowest, and release it when demand peaks in the hot summers. The mountainous nature of the plateau allows for deep reservoirs to be created in mountain valleys, cheaply holding large volumes of water and multiplying its value by increasing the potential energy that can be used to generate electricity for more people, but also withholding more fresh water from populations downstream.
This natural offset from seasonal demands, paired with the water being stored at high elevations, makes the plateau function as the ‘Water Tower of Asia’.
While India’s moves on the Indus have made the news, they do not come with the ability to fully implement India’s threat. Thanks to previous adherence to the Indus Water Treaty, New Delhi can only restrict 4.4 cubic kilometers of Pakistan’s rivers, far below the volume it would need to throttle the system for more than a few days – which is why the pact has historically withstood political shocks.
In contrast, Tibet’s waters are uniquely prone to being cut off in China. The rivers and basins of the Tibetan Plateau are clustered close together in a choke point known as the Three Parallel Rivers. At their narrowest point, the Mekong, Salween, and Yangtze Rivers are within about 32 kilometers of each other, with even shorter distances between basins. This means that if a dam or pump station were placed on the Mekong or Salween, these rivers’ flows could be diverted into the Yangtze basin, changing the endpoint of the flowpath by thousands of miles. In this way, even a small intervention sufficiently upriver could have an enormous impact on the flow balance of these complex systems.
Chinese dams on the Mekong and Brahmaputra Rivers in particular represent threats to the water security of Bangladesh and India, along with much of continental Southeast Asia. China’s planned Yarlung Zangbo hydro-electric project on the Brahmaputra River near the Indo-Chinese border is the flagship of Beijing’s hydroelectric ambitions. Once constructed, it will be three times larger than the Three Gorges Dam, or about the size of 25 percent of the entire existing Chinese hydroelectric infrastructure. On the Mekong, China has already built 11 large dams, and is constructing or planning at least eight more.
Across its hydro-portfolio, China plans to divert up to 200.6 billion cubic meters of water per year from the Tibetan Plateau to the Yellow and Yangtze river basins. This scheme diverts as much water as about five Grand Ethiopian Renaissance Dams. Were it not diverted, much of this water would flow out of China and into Southeast Asia and India.
The damming projects planned for these rivers go beyond diverting their course. The proposed dams also hold the water back. The downstream impact of these projects is already being felt: China’s damming activities have been linked to droughts on the Mekong River in 2019 and floods in Thailand in 2024.
By controlling when to divert flow or fill and empty their reservoirs, nations can threaten to reduce downstream countries’ access to water if they do not comply with their demands, with devastating consequences for the downstream economies and environments. On the flip side, countries could open the floodgates during the monsoon season, when areas downstream of the Tibetan Plateau are regularly inundated with local floodwaters.
India and Bangladesh are currently especially vulnerable due to their reliance on the Brahmaputra River for fishing, farming, transport, and industry. Bangladesh is a country built around a river delta, and any significant change to the Brahmaputra and Ganges Rivers would threaten the nation’s existence.
Because the history of water in China is so connected with the perception of a right to rule, the regime is also likely conscious of its need to maintain access to extremely cheap water for its own population as a means of preserving its legitimacy. Cheap water is also a subsidy to Chinese agriculture and industry, which further insulates the nation from foreign influence by making it more self-reliant and immune to sanctions.
In the last 20 years, the country has lost an estimated 28,000 of its 60,000 rivers to over withdrawal, pollution, and desertification – an area equivalent to the Mississippi River basin. Water scarcity is a looming problem for China, even in areas that have abundant water resources today. This helps explain why China is set on diverting so much water from the Tibetan Plateau: damming these rivers is a way to reserve for itself a bigger piece of a shrinking pie.
China has both the means and the method to accomplish its goals. The nation’s domestic infrastructure already diverts huge amounts of flow from the South of the country to the North. The South-to-North Water Transfer Project, which began in 2003 under President Hu Jintao, a former hydraulic engineer, transports nearly 50 billion cubic meters of water from the humid and flood-prone South to support agriculture and industry in arid major northern cities and provinces each year.
Much of this sprawling system of waterways, pipelines, and channels was built on the backbone of China’s Grand Canal. Construction on the Grand Canal dates to the fifth century BC and eventually resulted in a system spanning almost 1,800 kilometers. The Grand Canal connected the Yangtze, Yellow, and Huai rivers of China’s heartland and allowed people, goods, and water to travel between them without the high energy costs of overland travel. This hugely expanded the size of the Imperial Chinese market that the typical Chinese person could access and allowed for local shortages to be balanced by surplus across the empire.
The Grand Canal can be understood as a monument to the history of hydro-manipulation in China, when China’s ambitions for water hegemony were born. It dates back to a time when emperors and warlords sought divine right by the mandate of heaven. Their ability to provide water for irrigation and protect the farmlands and populace from floods was evidence of their right to rule. Mismanagement and disasters were seen not just as incompetence by the people, but as the renunciation of a ruler’s legitimacy by heaven itself.
The size of China’s present-day hydroelectric grid is similarly astounding: it produces 1,300 terawatt hours of energy each year, roughly equivalent to the output of all US coal and renewables combined. Last year, Chinese hydroelectricity produced about a quarter more energy than the entire continent of Africa.
This is not simply a product of China’s scale. Hydroelectricity’s share is disproportionate in comparison to the rest of the Chinese grid. Whereas the US grid runs on about five percent hydropower, the UK on about two percent, and Japan about seven percent, hydroelectricity accounts for almost 13 percent of total electricity generated in China. While there are other countries like Norway or Iceland that are even more reliant on hydropower, their small populations require far fewer projects to support.
So far, international policy has failed to curb the threat China poses to its downstream neighbors. The last major effort to govern China’s actions was a non-binding memorandum of understanding signed between Beijing and New Delhi in 2013. The memorandum recognizes that the countries share rivers, and should work together to manage them; it made no real promises and did not prompt any significant cooperation. Previous attempts by Southeast Asian countries to govern the Mekong, such as the 1995 Mekong Agreement, which would have created a multi-national governing body for the Mekong River, resulted in a political stalemate when China and Myanmar backed out.
Desalination: a possible solution
China’s economic ascendance has largely freed it from the sort of international pressure that might have driven it to sign such an agreement. In parallel, the increased domestic need for water makes further cooperative efforts even less likely to bear fruit.
Upstream nations have too much leverage to be convinced to share water resources more equitably. The solution might be to focus on reducing downstream nations’ reliance on those resources in the first place.
Desalination is the process of separating dissolved salt and water. Humans have been doing this for more than 8,000 years. But for most of history we’ve focused on harvesting the salt rather than the water, which we’ve simply let evaporate. If you’ve ever looked out the window of a flight into San Francisco Airport, you’ve likely seen this process at work, producing the salt that is used for industry, roads, and pharmaceuticals, as well as on your table.
Harvesting salt is as easy as filling a pond and waiting for the water to evaporate, while harvesting water is comparatively difficult. Seawater contains about 35 grams of dissolved salt per liter, and drinking water should contain no more than about 0.5 grams per liter of any dissolved materials, including salt, minerals, metals, and so on. This means we have to remove roughly 98.6 percent of the salt and every other contaminant present in water before it is fit for human consumption, which is a tall order.
There are two methods of desalination in use in the roughly 18,000 desalination plants around the globe: distillation and reverse osmosis.
Distillation is a contained version of the normal water cycle. Salt water is heated into steam, which is then cooled, condensed, and captured as freshwater distillate. A great deal of energy is needed to initiate the phase change from water to steam. But these costs can be offset by using waste heat from power plants and avoiding scale buildup from deposited salts.
Flash evaporation is a type of distillation that manipulates pressure to boil different parts of a mixture at different times, allowing for separation. It is highly efficient and largely avoids the problem of scale buildup, thanks to the low temperature at which it operates. Flash evaporators are especially popular in freshwater-scarce environments like the Middle East and oceangoing ships because they are so efficient and reliable. Aircraft carriers commonly use flash distillation because their onboard nuclear reactors produce ample spare power for the energy-intensive process.
The second approach is reverse osmosis, which is the most intense form of membrane-based desalination. It uses high-pressure pumps to push saltwater against a membrane that has pores small enough for the water particles to pass through, but not the salts. This separates the water from the salt and other dissolved solids. As with distillation, the main problem with reverse osmosis is the energy cost. Running the pumps is expensive and energy-intensive. The size of the particles in the water is another challenge. Dissolved salts are among the smallest particles in water filtration, which means the pores in the membrane need to be extremely small. And the smaller the pore, the higher the pressure needs to be to push the water molecules through. Fortunately, advancements in membrane-based desalination technologies have reduced the cost of fresh water substantially, making desalination a viable option for many countries that were previously unable to afford it.
With the advent of cheap solar power, the last chokepoint that has frustrated global desalination efforts is finally loosening. The cost of the energy used by desalination plants over time dwarfs the initial cost of their construction, so the availability of cheap solar power greatly increases their viability. The ease with which countries can store water also means that solar power’s chief drawback, intermittency, may be much less of a problem here than in other uses.
Of China’s downstream neighbors, only Laos lacks access to a coastline from which to withdraw saltwater. But even they may be able to cooperate with their coastal neighbors to manage their shared rivers.
Downstream countries don’t necessarily need sufficient treatment capacity to fulfill their entire annual demand. It might only be necessary to be able to create a reserve of freshwater that is large enough to cover demand during the dry season, so that these nations can resist pressure precisely when they are most vulnerable. Having access to cheaper membranes for reverse osmosis also opens up opportunities to treat and reuse fresh water more efficiently.
These countries could also build and fill their own freshwater reservoirs to empty into existing rivers, or use technologies like aquifer storage and retrieval, in which a bubble of clean water is pumped into underground reservoirs and held for later, to store clean, membrane produced or distilled water in local aquifers where it won’t evaporate. Existing intakes on the river can then pick up the water and deliver it to residents as they always have. This means that countries won’t need to build an entirely new water distribution infrastructure as long as they can build the pipelines to transport water from the desalination plants to the existing river basins.
One way to reduce the costs involved in reverse osmosis would be to make the process more efficient by only treating seawater to the point that it can be mixed with the water the nation already has access to. Water produced by reverse osmosis is typically much purer than is needed, and is often intentionally remineralized. Allowing a certain amount of dilution with groundwater, for example, would be less costly while still producing water that is safe to drink.
Rivers are now battlefields
The threat of water scarcity is a global phenomenon on the rise. Like many nations around the world, the countries of southeast Asia are dependent on water controlled by an upstream neighbor. As more dams are constructed on the headwaters of rivers like the Indus, Mekong, and Brahmaputra, both the threat of water scarcity and the prospect of becoming tributary states loom large over the region’s countries.
The most obvious example of this hydro-political arms race expanding beyond China is India’s suspension of the Indus Water Treaty. Although India lacks the infrastructure to wholly follow through on the action, it shows that rivers are now battlefields.
Of greater concern is when capabilities coincide with intent. As recently as July 19th 2025, China began the construction of a megascale dam on the Brahmaputra river, which India and Bangladesh strongly oppose. It is estimated that the new dam will surpass the scale of the famous Three Gorges Dam by more than a factor of three, making it the world’s largest hydropower dam.
Desalination can turn the zero-sum game around transboundary freshwater resources into a positive sum game. In the near term, desalination technology can provide downstream countries with a reliable source of freshwater, reducing their dependence on upstream countries and making them less vulnerable to pressure. Over the long term, it has the potential to increase the freshwater supply for all nations, neutralizing much of the world’s water-based conflict.
Connor Tabarrok is an engineering consultant at Kimley-Horn and Associates and the author of the Substack Of All Trades. Follow him on Twitter
 | A guest post by
|
The description of the Yarlung Tsangpo project as 'three times larger than Three Gorges' is conventional in a sense but also misleading. It's only 3x larger by power capacity, but it's vastly smaller by flow rate and reservoir size. The project neither intends, nor will be capable, of impounding significant quantities of water or significantly disrupting the flows on the lower Brahmaputra. It really isn't comparable with the Ethiopian GED or the Mekong damsf You gloss over this and I think it would be better to be clear.
The hydropower equation is — of course — flow x head, and nearly all the power in the project is coming from the incredibly large head (nearly 2.5km). It's a series of large sluices with four very long tunnels, in essence. The sluices are unremarkable except for their remoteness. It's the tunnels which are unprecedented in size. There's no need for a vast reservoir because the topography does all that work already.
There are a series of narratives around this project which are obviously politically motivated. Much of the anti-campaign has already switched to arguments about sediment loss which also seem pretty speculative. I think WIP should try to be clearer about the technical facts although it would mean giving up your clickbait headline!
Very interesting article. I'd love to read a good piece on the economics of water desalinisation. I'm from Portugal where every summer there's talk of water scarcity, however, many argue it would be more cost efficient to transfer water from the north of the country to the south. I have my doubts given the abundance of solar energy and what has been achieved in places like Israel and the UAE.