Transboundary water disputes concern more than the division of a river. Dams also distribute the power to change its seasons, the information needed to anticipate those changes, and the risks borne by people who depend on water arriving at the right time.
On 23 March 2014, water was released through Morelos Dam into the Colorado River Delta in Mexico. The release formed part of an eight-week environmental pulse negotiated under Minute 319, an agreement between the United States and Mexico. Water normally managed chiefly for farms and cities was being deliberately supplied to a damaged river corridor. The event required cooperation between governments, scientists and conservation organisations, followed by the less ceremonial work of measuring what the release achieved. Opening the gates was an identifiable act. Its consequences spread unevenly through the landscape below them.
As the pulse advanced, water entered dry ground, raised groundwater levels and encountered channels whose vegetation and physical condition affected its passage. Much of the released volume infiltrated rather than continuing downstream as surface flow. Infiltration could support groundwater and vegetation, but it also complicated delivery to particular restoration sites. The monitoring programme exposed a distinction that diplomatic agreements can obscure: a volume released from a structure and a useful flow arriving at its destination are different things. Delivery depended on the channel, the route and the ecological purpose of the release.
The Colorado experiment belongs to a larger political history. Across shared rivers, governments negotiate quantities while communities depend on patterns: a flood that arrives during a growing season, a reservoir that survives a dry year, a release announced early enough for someone downstream to prepare. Transboundary water disputes become difficult partly because these needs cannot all be expressed through an annual total. A dam gives its operator choices over some of those patterns. The public significance of that power lies in who can influence the choices, establish their effects and secure a change when the arrangement causes harm.
What Water Carries
A reservoir stores water across time. Holding part of a seasonal flood can make it available during a dry month, generate electricity when demand is high or leave space for an anticipated flood. These functions can support households and productive investment, but they also introduce competing preferences over the river’s behaviour. Farmers may need irrigation at one point in the agricultural calendar; electricity buyers may value generation at another. Operators work within physical constraints, including storage capacity, incoming flows, outlet capacity and safety requirements. A large storage reservoir and a plant with little capacity to retain water offer very different degrees of discretion.
Hydroelectric generation generally passes water downstream through turbines, yet storage changes the timing of its passage and introduces evaporative losses. Electricity production also depends on the height through which water falls, giving operators reasons to preserve reservoir levels. Commercial arrangements enter the calculation through demand, delivery commitments and the revenues needed to sustain the project. A dam’s promised output must eventually find buyers and usable access to the electricity network. An upstream location therefore supplies a physical opportunity whose political value depends on finance, markets, institutions and the ability to keep the installation working.
The water itself carries more than volume. Sediment transports nutrients, helps form habitats and contributes to the maintenance of floodplains and deltas. Seasonal rises can connect a river to the surrounding landscape, supporting activities that depend on periodic inundation. Smoothing a flood may protect one settlement while changing the conditions on which another livelihood rests. The Mekong River Commission’s sediment monitoring work identifies pressures from sand extraction, land-use change, irrigation and groundwater pumping, alongside dams. Assessing a river’s deterioration requires separating these interacting influences. A report showing that sufficient water passed a gauge can leave important questions unanswered about what that water carried and what functions it could still perform.
The Inherited Rules of Transboundary Water Disputes
The Colorado’s governing arrangements illustrate how earlier decisions acquire lasting authority. The 1922 compact established a division between its Upper and Lower Basins, while the 1944 treaty created delivery obligations to Mexico. Federal law, state water rights, court decisions and later agreements added further layers. Tribal water rights complicate any account that treats the basin as a negotiation among seven states alone: their origins and priority can differ from those of other users, and recognition of a right does not automatically provide the infrastructure and arrangements needed to exercise it. Disputes over shortage consequently take place within an inherited distribution of claims and institutional influence.
These are different legal relationships. An interstate compact within the United States operates inside a federal constitutional order; an international treaty between the United States and Mexico joins separate sovereign states. Treating both simply as disagreements between upstream and downstream users would erase the institutions through which claims can be enforced. Their comparison is nevertheless revealing. An existing use can become socially indispensable long before the people who depend on it encounter the limits of the arrangement that made it possible. Revising that arrangement then means changing expectations around which farms, homes, public budgets and investments have already been organised.
The Nile carries a different inheritance. Egypt and Sudan’s 1959 agreement allocated water between them and provided for major works, including the High Aswan Dam. Ethiopia was not a party to that bargain. The agreement shaped infrastructure and expectations, while its bilateral origins matter to the claims made by other riparians. Protecting established livelihoods is a substantial political obligation. It also raises a difficult question when the protection sought would constrain development elsewhere: how much authority should an earlier arrangement exercise over populations that had no part in making it? New infrastructure can challenge an inherited privilege while creating dependencies of its own.
In March 2015, Egypt, Ethiopia and Sudan signed a Declaration of Principles specifically concerning the Grand Ethiopian Renaissance Dam, or GERD. Principle V linked joint studies to agreement on rules for initial filling and annual operation. It also envisaged adjustments by the dam’s owner, notification of unforeseen or urgent circumstances requiring operational changes, and a mechanism for coordination with downstream reservoirs. The declaration supplied a shared framework for developing rules without itself prescribing a complete operating schedule. Its wording brought the central problem into view: how to preserve an operator’s ability to respond to changing conditions while making the consequences more predictable for others.
International water law offers principles for assessing competing claims without producing an automatic winner. The 1997 UN Watercourses Convention combines equitable and reasonable use with duties to take appropriate measures to prevent significant harm and to cooperate. Its assessment factors include existing and potential uses, social and economic needs, effects on other states and available alternatives. Treaty participation, basin agreements and other applicable international law determine the obligations relevant to particular states. Neither the age of a use nor a country’s position upstream settles the inquiry. Research on hydro-hegemony similarly examines how bargaining, institutions, knowledge and material resources shape control. Geography influences those relationships; it does not assign every upstream state dominance or every downstream state helplessness.
After the Drought Ends
On the Nile, even a technically informed argument becomes misleading if it compresses the life of a reservoir into one confrontation over filling. Initial impoundment, ordinary operation, prolonged drought and recovery require different decisions. Ethiopia’s generation objectives, Egypt’s reliance on stored supplies and Sudan’s irrigation, reservoirs and exposure to changing Blue Nile flows intersect differently in each phase. Sudan has interests in operating the system, rather than merely a position between two larger arguments. Understanding the basin requires following what each country needs the river to do as storage rises or falls.
A 2020 study in Nature Communications modelled these interactions using specified operating assumptions and historical sequences of flows. It identified potential Sudanese benefits from steadier flows, including irrigation and reductions in damaging floods and reservoir sedimentation. It also showed how releases for generation could help reduce downstream shortages during part of a severe drought. Those were conditional results from a simulation, not observations establishing what subsequently happened. In the model, depletion eventually limited the protection available from both the GERD and the High Aswan Dam.
The recovery phase introduced another conflict. After a long drought, improved inflows would allow both reservoirs to rebuild storage. The speed and sequence of refilling could affect downstream supplies and the restoration of Ethiopia’s generation capacity. Rain returning to the basin would not, by itself, settle those choices. Recovery has a distribution of its own: one operator may seek to rebuild a safety margin while another still needs releases to sustain current use. A settlement covering the emergency but leaving recovery undefined can defer part of the disagreement until the moment when the weather appears to have improved.
This is a demanding test for agreements advertised as mutually beneficial. Benefits during ordinary years can encourage parties to postpone the rules needed for exceptional ones. Each government may prefer discretion until another government’s exercise of discretion becomes threatening. Advance commitments exchange some flexibility for greater predictability, and the value of that exchange varies with exposure and bargaining power. Examining how an agreement works when reservoirs are depleted—and how it governs their replenishment—reveals obligations that cooperation in average conditions may leave untested.
Knowing the River, Knowing the Plan
Much of the political power surrounding a dam resides in information. A downstream agency may know the water level at a gauge yet remain uncertain about the release planned for tomorrow. A seasonal forecast may arrive early enough to guide planting, or too late to change anything. Satellite observations can help establish reservoir conditions without revealing operational intentions. Measurements made after a damaging event can support investigation while offering little protection beforehand. A claim that data have been shared therefore needs to specify which data, at what interval, with what warning and for whom.
The Mekong demonstrates why that assessment needs an accurate starting point. Cambodia, Laos, Thailand and Vietnam concluded the 1995 Mekong Agreement. Cooperation with China and Myanmar developed through a separate dialogue process, launched in 1996. China began supplying year-round hydrological data from two stations in November 2020. The record of cooperation establishes an information channel whose coverage and usefulness can be examined. Observed flows can improve forecasting while leaving further needs for information about planned releases, operational changes and the circumstances under which earlier plans might be revised.
Attribution presents a separate difficulty. The Mekong River Commission’s assessment of low flows in 2019–2021 examined rainfall, storage and reservoir operations together. Reduced rainfall could lower reservoir inflows, influence operators’ decisions and diminish the water subsequently available for release. A low reading downstream cannot establish, on its own, either the contribution of a particular dam or an intention to cause damage. Establishing an operational contribution also does not exhaust responsibility: the availability of alternatives, prior warnings and responses to emerging harm still matter. An investigation needs to make these distinctions clearly enough for uncertainty to be examined without bringing scrutiny to a halt.
The same report makes the importance of timing unusually concrete. During the wet season, higher Mekong levels send water back towards Cambodia’s Tonle Sap Lake, expanding its flooded area. Releases to support the timing of that reversal could be needed when reservoir storage was low, creating risks for power production; releases aimed at improving the eventual volume might be more feasible later. Restoring an annual quantity and restoring a seasonal function are different undertakings. An agreement can record an apparently adequate volume while the dispute persists over when it becomes useful.
A Consultation That Changed the Dam
Legal procedures attempt to bring such concerns into decisions before construction makes revision expensive. The Mekong Agreement distinguishes notification, prior consultation and specific agreement according to location, season and type of use. Its definition of prior consultation gives neither a veto nor an unrestricted unilateral entitlement to proceed without considering other riparians’ rights. Objections must have consequences while a process capable of permitting action remains available. Counting meetings cannot resolve that tension. The relevant evidence lies in what participants could examine, what they requested and what happened to those requests.
The Xayaburi project in Laos provides a record that supports a more discriminating judgment than either celebration or dismissal. Prior consultation began in October 2010. A 2019 technical review by the Mekong River Commission Secretariat recorded substantial redesign and investment responding to concerns, including changes to fish-passage facilities and the addition of low-level gates intended to help sediment movement. It also found that important information, including operational details and baseline data, was insufficient for a full assessment of the likely effectiveness of several measures. At that stage, responsiveness coexisted with unresolved questions about protection.
Those findings identify three achievements that should be judged separately: considering an objection, changing an installation and demonstrating that the change works. A revised engineering drawing establishes a decision to alter a project; protection of a fishery requires evidence about a functioning river over time. Equally, gaps in the review material do not prove that every modification failed. Xayaburi shows how consultation can influence a project while leaving an unfinished obligation to establish results. The date and scope of the review matter: a concern recorded in 2019 cannot serve as a verdict on all subsequent operation.
Operating rules connect the engineering changes to their use. A gate’s dimensions establish a capability, while the timing and conditions of opening it determine what that capability achieves. Monitoring therefore needs procedures for changing operations when results fall short, with responsibilities and resources assigned in advance. Otherwise, a technically serious review can identify a problem that nobody has a sufficiently clear mandate to correct. Participation gains practical force when affected parties can bring evidence into that process and obtain a response.
The National Benefit and the Local Loss
International negotiations often speak of each country as a single recipient of benefits. Inside it, electricity consumers, treasury officials, project owners, displaced households and fishing communities can occupy very different positions. Lower Sesan 2 in Cambodia, a tributary project distinct from Xayaburi, illustrates this internal distribution. Human Rights Watch’s 2021 investigation identified China Huaneng Group as the controlling investor and reported the displacement of nearly 5,000 people, mostly Indigenous people and other ethnic minorities. Drawing on interviews with more than sixty people, it documented accounts of damaged livelihoods and inadequate compensation, while examining official and company claims about the project’s benefits.
One interviewee, identified by the pseudonym Dao, described a livelihood combining fishing, farming and forest products. He rejected the relocation offer and built another home outside the flood zone, yet reported losing much of the income those activities had supplied. His account exposes a problem that a replacement-property valuation can miss: a household may depend on resources dispersed across its surroundings. Some losses can be addressed through money, assets or services; others require continuing access to a resource or restoration of a function. Compensation must be assessed through what households can sustain afterwards, including those who accepted relocation, resisted it or found another place to live.
That distribution changes the meaning of a project’s national success. Additional generation can be valuable while a community loses the means of sustaining itself. Governments have legitimate reasons to build infrastructure serving populations far beyond a project site. Their case requires an explanation of how concentrated costs will be addressed, whether alternatives are workable and what happened to people after relocation. An aggregate benefit cannot answer those questions by itself, even when the aggregate calculation is sound. The people included in a national total may possess sharply unequal access to its gains.
The distribution also extends through trade. Britain’s 2024 Food Security Report identifies reliance on food imports from countries under greater water stress, including Spain and Morocco. Consequences depend on production, prices and alternative suppliers, so a local shortage does not translate mechanically into the same loss abroad. Importers may be able to change suppliers or buy substitutes; a fishing community may remain tied to a particular river and season. Water security crosses borders through the food economy as well as through river channels, connecting consumers with decisions whose local consequences they may never see.
Rules for the Years Nobody Wants
Shared rivers require decisions at speeds that exhaustive renegotiation cannot accommodate. Reservoir operators responding to a flood cannot wait for a fresh diplomatic settlement over every gate movement. Governments also have to plan before hydrological forecasts become certain. In transboundary water disputes, delegating operational authority therefore requires rules about its limits. Predetermined drought thresholds, warning obligations and procedures for departing from normal releases can preserve the capacity to act. Their design remains political: each threshold distributes risks, and the definition of an emergency can alter whose needs receive priority.
The triggers themselves need scrutiny. A rule based only on the level of one reservoir may respond to earlier withdrawals and operating choices as well as to the weather. Information from across the system can help distinguish those influences. Rules for recovery can prevent the first improved inflows from becoming an unregulated contest to refill. Agreements also need procedures for responding when monitoring shows that an agreed protection is failing, including how the costs of adjustment will be handled. An objection supported by evidence must be able to reach a decision before the interest at stake has been irretrievably damaged.
The Colorado’s environmental pulse shows why the duration of a benefit belongs in that assessment. Monitoring recorded increased vegetation greenness in 2014. By 2017, however, that measure had returned to or fallen below its 2013 level in most river reaches, while remaining higher in others. In one of the reaches with more persistent gains, the report identified restoration activities, including continuing base flows, as a possible explanation. Those findings qualify the story of a river revived by opening a gate. They direct attention to what sustained particular sites after the pulse had passed.
The experience supports agreements that can learn from delivery. Routing, duration, groundwater conditions and habitat work affected what environmental water could achieve. A negotiated allocation created an opportunity; repeated observation helped establish where that opportunity had been realised and where further work was needed. Cooperation becomes more credible when institutions can explain how those observations changed the next decision, including what remained beyond the intervention’s reach.
At Morelos, governments could agree on a release and engineers could open the gates on schedule. The water still had to travel through a channel shaped by earlier withdrawals, vegetation and years of altered flow. Restoring particular sites required attention to that accumulated history. Every shared river joins decisions made at these different scales: a treaty, a reservoir schedule, a gauge, a field, a fishing ground. The measure of shared governance is whether evidence from those places can change the next release while there is still time for that change to matter.
