The Role of International Water-Sharing Agreements in Revising Resource Conflict

The global freshwater situation has a structure that makes conflict its default outcome and cooperation its achievable alternative. Of the world's approximately 310 major river basins, 276 are transboundary — shared by two or more nations. Of the world's major aquifers, about 300 cross international borders. The world's major river systems — the Nile, the Mekong, the Tigris-Euphrates, the Amazon, the Congo, the Colorado — collectively drain basins containing the majority of the world's agricultural production and a large fraction of its population.

The default logic in this situation is Hobbesian. Upstream nations can extract water before it reaches downstream nations. Nations with greater military capacity can enforce their preferred allocation. Historical use claims can conflict irreconcilably with current need. Development disparities create asymmetric pressures: a developing nation may need water for agriculture and industry that a developed nation is using for less essential purposes but has established legal priority to. The interaction of these forces, without governance frameworks, tends toward conflict.

The historical record confirms this tendency. The term "rival" derives from the Latin rivalis — one who shares a river. Water conflicts are among the oldest documented forms of human warfare. The Pacific Institute's Water Conflict Chronology database documents over 900 water-related conflicts across human history, with the frequency of conflict increasing significantly since 2000.

But the same database also documents a different phenomenon: the extensive development of cooperative water agreements, many of which have prevented or resolved conflicts that water stress would otherwise have generated. This coexistence of conflict and cooperation in water governance is the central phenomenon that international water law attempts to understand and promote.

The Architecture of Water Agreements

International water agreements have developed a characteristic architecture over the past century, refined through practice and failure into a set of elements that functional agreements tend to share.

The first element is quantified allocation. Vague commitments to "equitable use" or "no significant harm" are insufficient for operational management. Functional agreements specify what each party is entitled to withdraw, in what form (surface flow, groundwater), in what timeframe (annual, seasonal), and under what conditions (average flow years, drought years). The Nile Waters Agreement of 1959 between Egypt and Sudan allocated specific volumes — 55.5 billion cubic meters to Egypt, 18.5 billion to Sudan — leaving 10 billion for evaporation losses and no provision for the other nine Nile riparian nations, which is one reason it has been politically unstable.

The second element is a monitoring mechanism. An allocation that cannot be monitored cannot be enforced. Functional water agreements establish joint technical bodies with access to hydrological data, withdrawal records, and physical monitoring infrastructure. Remote sensing technology has dramatically improved monitoring capacity over the past two decades: satellite imagery can detect changes in reservoir levels, surface water extent, and irrigation withdrawal patterns with sufficient precision to identify material violations without requiring physical access to the other party's territory.

The third element is a dispute resolution mechanism. Even well-designed agreements will generate disputes about interpretation, measurement, and exceptional circumstances. Agreements that specify the dispute resolution procedure — from technical consultation to diplomatic negotiation to arbitration to international legal adjudication — are more stable than those that leave dispute resolution unspecified, because parties know in advance how disagreements will be addressed and are less likely to escalate unilaterally.

The fourth element is a revision mechanism. This is the element most often absent from historical agreements and most consequential for long-term stability. Static agreements, made to reflect conditions at a single historical moment, become increasingly misaligned with reality as hydrology, population, technology, and political relationships change. Agreements with built-in revision procedures — scheduled reviews, amendment processes, renegotiation triggers linked to defined conditions — can adapt without requiring the parties to abandon the agreement entirely and renegotiate from scratch.

The Indus Waters Treaty: A Detailed Study

The Indus Waters Treaty is the world's most extensively studied international water agreement, both because of its longevity (signed in 1960, still in force sixty-plus years later) and because of its extraordinary context: it has governed the water relationship between two nuclear-armed nations that have fought four wars (1947, 1965, 1971, 1999) and have been in a state of near-continuous political tension for their entire existence as independent states.

The treaty was brokered by the World Bank following eight years of negotiations. Its basic structure is a territorial division: the three eastern rivers (Ravi, Beas, Sutlej) flow primarily to India; the three western rivers (Indus, Jhelum, Chenab) flow primarily to Pakistan. Each country receives the waters of its allocated rivers; the other country has limited rights to use the opposite rivers for specific, defined purposes (non-consumptive uses, limited agricultural use, specified hydropower development).

The treaty's durability under extraordinary political stress has been the subject of substantial academic analysis, and several structural features of the treaty are consistently identified as explanatory:

First, the clean division principle. Rather than requiring the two countries to jointly manage a shared river system, the treaty divided the system into components, with each country having primary control over its component. This reduced the need for ongoing cooperative management — which was always going to be difficult given the political relationship — while establishing clear boundaries within which each country could develop its water resources independently.

Second, the Permanent Indus Commission. Two commissioners, one from each country, meet annually and communicate continuously. Their mandate is technical rather than political: data sharing, notification of planned works, joint inspections, and first-level dispute resolution. The commission has maintained communication and met its annual meeting schedule continuously, including during periods of active military conflict. The separation of technical water management from political relations has proved to be a genuine institutional achievement.

Third, the detailed technical annexes. The treaty includes extensive annexes specifying exactly what each country is and is not permitted to do with the rivers — agricultural withdrawal limits, hydropower development specifications, storage restrictions. This level of technical detail reduced interpretive ambiguity in ways that vaguer agreements do not.

The treaty has nonetheless been under stress in recent decades. India's hydropower development on the western rivers — particularly the Baglihar Dam and the Kishenganga hydropower project — has generated sustained disputes from Pakistan about whether the projects comply with the treaty's restrictions. These disputes have proceeded through the treaty's dispute resolution mechanisms, with some going to international arbitration under Annex G of the treaty. The fact that disputes are being resolved through treaty mechanisms rather than through unilateral action or military threat is itself a measure of the treaty's success, even when the outcomes are contested.

Climate change is creating a new stress category. The Indus basin's hydrology is heavily dependent on glacial melt from the Himalayas and Karakoram ranges. Climate change is accelerating glacial retreat, which is initially increasing summer river flows but will eventually reduce them as glacial mass is depleted. The treaty was negotiated under hydrological conditions that are now changing irreversibly. Neither party has invoked the treaty's provisions for this scenario, because the treaty was not designed with it in mind. The pressure for treaty revision is building but has not yet produced a formal renegotiation process.

The Nile: The Limits of the Model

The Nile basin offers the counterexample — a water governance situation where the existing frameworks have been inadequate and conflict potential remains high.

The Nile is shared by eleven nations. Egypt and Sudan concluded the 1959 Nile Waters Agreement dividing the river's flow between them, without consulting the nine upstream riparian nations. Ethiopia, which provides approximately 85 percent of the Nile's total flow through the Blue Nile drainage system, was not party to the agreement and has never accepted its validity.

For decades, the de facto governance was maintained by Egyptian political power — Egypt threatened military action against any upstream dam construction that would reduce its share — and by the limited development capacity of upstream nations, which lacked the capital and technology to develop their water resources significantly. The Nile Basin Initiative, launched in 1999, created a cooperative framework among the riparian nations for the first time, but it remained advisory rather than legally binding.

Ethiopia's construction of the Grand Ethiopian Renaissance Dam (GERD), begun in 2011 and reaching initial operation in 2020, brought the underlying governance failure into acute crisis. GERD is the largest hydropower dam in Africa and significantly affects downstream flow, particularly during the filling period. Negotiations among Ethiopia, Egypt, and Sudan — facilitated variously by the African Union, the United States, and other international parties — have repeatedly failed to reach a binding agreement on filling schedules, operating rules, and drought management.

The GERD negotiations illustrate several failure modes in international water governance. First, the absence of a pre-existing binding framework covering all parties: the 1959 agreement that Egypt treats as authoritative, Ethiopia treats as having no legal force, since Ethiopia was not a party. Second, deep power asymmetry within a context of shifting relative power: Egypt was historically dominant but Ethiopia is developing rapidly and cannot accept Egyptian veto over its infrastructure decisions. Third, the absence of trust-building mechanisms: unlike the Indus Commission, there is no technical body with a sustained relationship and a mandate for data sharing and joint analysis. Fourth, the absence of third-party arbitration authority that all parties accept: without a dispute resolution mechanism with binding authority, negotiations have no resolution mechanism when bilateral talks fail.

The GERD situation has not produced armed conflict, partly because both Egypt and Ethiopia have more pressing domestic challenges, and partly because international diplomatic pressure has kept channels open. But it represents a water governance failure in one of the world's most water-stressed regions, and the unresolved nature of the dispute will generate increasing pressure as climate change reduces total Nile flow and agricultural water demand increases.

The Colorado River: A Domestic Collapse in Progress

The Colorado River Compact of 1922 divided the Colorado's flow among seven U.S. states based on a measurement period that happened to be unusually wet. The compact allocated 17.5 million acre-feet annually from a river whose long-term average flow was closer to 14.8 million acre-feet. The overallocation was built in from the beginning.

For decades, the overallocation was masked by groundwater extraction, which compensated for surface water deficits, and by reservoir storage, which carried surplus wet-year flows to cover dry-year deficits. Lake Mead and Lake Powell, the two major Colorado storage reservoirs, provided a buffer of years of stored water.

By 2021, that buffer had been nearly exhausted. Lake Mead fell to 27 percent of capacity — the lowest level since it was filled — and Lake Powell reached emergency levels. The Bureau of Reclamation declared the first-ever Tier 1 shortage, triggering mandatory cuts to Arizona's water allocation. Tier 2 and Tier 3 shortage declarations followed, with cuts expanding to Nevada and potentially to California.

The Colorado crisis is a model of what happens when a water agreement lacks adequate revision mechanisms. The 1922 Compact has no systematic review procedure. Its allocation numbers have never been formally revised to match actual hydrological conditions. The political difficulty of reducing allocations — which requires states to accept reduced entitlements they have built infrastructure and economies around — has prevented the revision that the physical reality has demanded for decades.

The crisis has forced improvised negotiations. The seven basin states, under federal pressure, have agreed to voluntary temporary cuts — the Drought Contingency Plans of 2019 and subsequent agreements — but these are emergency patches, not fundamental revisions. The underlying compact remains unamended, with allocations that exceed available water by several million acre-feet annually, a gap that climate change is widening.

The Colorado situation illustrates a general principle: water agreements without revision mechanisms eventually fail when conditions change, and the absence of procedural revision forces the crisis-driven revision that damages institutions and trust in ways that procedural revision does not.

Climate Change as the Stress Test

Every major water-sharing agreement currently in force was designed for a hydrological reality that climate change is now departing from. Reduced snowpack in mountain watersheds, accelerated glacial retreat, shifting precipitation patterns, increased evaporation from higher temperatures, and more frequent extreme events — droughts and floods both — are changing the water availability that agreements allocate.

The International Water Law Project and related academic institutions have identified the specific legal and institutional challenges this creates:

First, agreements that allocate fixed volumes will face increasing pressure as those volumes become unavailable in dry years. Percentage-based allocations — which allocate shares of whatever flow exists — are more climate-resilient than volume-based allocations, but most historical agreements use volume.

Second, agreements that define drought triggers based on historical frequency distributions will be systematically wrong as those distributions shift. What was a one-in-twenty-year drought may become a one-in-five-year drought, triggering emergency provisions far more frequently than the agreement's designers expected.

Third, aquifer-based agreements are particularly vulnerable because aquifer recharge rates are highly climate-dependent and frequently unknown with sufficient precision. Agreements that permit withdrawal rates based on assumed recharge levels may be permitting overdraft that is invisible until the aquifer begins to decline.

Fourth, upstream-downstream relationships change character under climate stress, because the dynamics of glacier-fed rivers (initially high flow from accelerated melt, eventually collapsed flow as glaciers disappear) are qualitatively different from the stable seasonal patterns on which agreements were designed.

The Civilizational Stakes

Water stress is projected to affect two-thirds of the world's population within the current century. The regions under greatest stress — South Asia, the Middle East, North Africa, sub-Saharan Africa, Central Asia — are also among the most politically fragile and the most consequential for global stability.

The revision from water-as-conflict-trigger to water-as-cooperation-domain is not a single event. It is a continuous process that requires sustained institutional investment, diplomatic engagement, and political will. The agreements that have worked — the Indus Waters Treaty, the Rhine Action Programme that cleaned up the heavily polluted Rhine in the 1980s, the Columbia River Treaty between the United States and Canada — have worked because they combined clear initial allocations with genuine monitoring capacity and institutional mechanisms for ongoing adjustment.

The agreements that have failed or are failing have failed because one or more of these elements was absent: because allocation was imposed rather than negotiated, because monitoring was inadequate, because revision mechanisms were not built in, or because the underlying power relationships made compliance optional for the more powerful party.

The civilizational task is to build the governance infrastructure — legal frameworks, monitoring systems, institutional bodies, dispute resolution mechanisms, and revision procedures — that converts water from a source of war to a framework for cooperation at a scale sufficient to address the water stress that climate change is generating. This is not utopian. It has been done before, in specific basins and contexts. The question is whether the pace of institutional development can match the pace of hydrological stress — whether the revision of governance can keep ahead of the revision of climate.

On the answer to that question, a significant fraction of civilizational stability in the twenty-first century depends.