The Colorado runs about 2,330 km from the Rocky Mountains of northern Colorado toward the Gulf of California, draining roughly 640,000 km² across seven US states and two Mexican ones. It supplies water to something like 40 million people and, in most years, does not reach the sea — because the agreements dividing it up allocate more water than the river actually carries.
A river that cuts rather than meanders
The Colorado begins as snowmelt in the Rockies and spends most of its length on the Colorado Plateau, a broad uplifted block of layered sedimentary rock. That combination — a river with a steep gradient and a heavy sediment load, crossing rock that was rising slowly beneath it — is what produced the canyon country of the American southwest.
The Grand Canyon is the famous result: roughly 446 km long, up to about 29 km wide, and over 1,800 metres deep, exposing rock layers spanning a very large part of Earth's history. But it is one of many. Glen Canyon, Cataract Canyon, Marble Canyon and the tributary gorges of the Green and San Juan rivers are all products of the same process.
Before the dams, the Colorado was extraordinarily silty — the name is Spanish for "coloured" or "reddish", after the sediment. It also ran wildly unevenly, with spring snowmelt floods an order of magnitude above late-summer flows. Both characteristics have been almost entirely removed by engineering.
The 1922 Compact and the number that was wrong
The Colorado River Compact, signed in 1922, divided the river between an Upper Basin (Colorado, Wyoming, Utah, New Mexico) and a Lower Basin (California, Arizona, Nevada), allocating each 7.5 million acre-feet per year, with a later treaty granting Mexico 1.5 million.
The problem is the figure it was built on. Negotiators used flow records from the preceding two decades, which tree-ring reconstructions have since shown to have been among the wettest periods in centuries. The long-run average flow is materially lower than the compact assumed — and the total allocated, once Mexico's share and evaporation losses are included, exceeds what the river delivers in a typical year.
For decades this was concealed by the Upper Basin not using its full entitlement and by large reservoirs absorbing the difference. Neither cushion is available now. Extended drought since 2000, compounded by higher temperatures increasing evaporation and reducing snowpack, has drawn Lake Mead and Lake Powell down to levels that have repeatedly approached the elevations below which their dams cannot generate power.
The result is a slow-motion negotiation among seven states, Mexico, and thirty tribal nations with senior water rights, over which uses get cut. Roughly three-quarters of the river's consumptive use is agricultural, and a substantial share of that grows livestock feed — which is why proposals to reduce alfalfa acreage recur, and why they are politically explosive in irrigation districts whose entire economy rests on that water.
Two dams that changed everything downstream
Hoover Dam, completed in 1936 in Black Canyon, was the largest concrete structure in the world when built and created Lake Mead, the largest reservoir in the United States by capacity. It made large-scale settlement of southern Nevada and much of southern California practical, and it ended the annual flood cycle on the lower river.
Glen Canyon Dam, finished in 1963 upstream of the Grand Canyon, created Lake Powell and was more controversial — it flooded a canyon system that relatively few people had seen, and its construction is often cited as a formative moment for the American environmental movement.
Their combined effect on the river below is profound. Sediment settles out in the reservoirs, so the water released into the Grand Canyon is clear and sediment-starved, which erodes the sandbars that beaches, backwaters and native fish habitat depend on. Water released from deep in the reservoir is cold and of constant temperature year-round, which suits introduced trout and disadvantages native warm-water species like the humpback chub and the Colorado pikeminnow. Managers now conduct deliberate high-flow experimental releases to rebuild sandbars — engineered floods standing in for the natural ones the dams removed.
The delta that dried up
The Colorado once ended in a delta of roughly 8,000 km² at the head of the Gulf of California — a wetland of cottonwood and willow forest, brackish lagoons and enormous bird populations, described by Aldo Leopold in the 1920s as a green lagoon of extraordinary richness.
Upstream diversion has reduced it to a fraction of that. For most years since the 1960s the river has been fully consumed before reaching Mexico's coast, and the delta has become largely a salt flat, with remnant wetlands sustained mainly by agricultural return flows. The Gulf's upper reaches lost the freshwater input their fisheries depended on, a factor in the collapse of the totoaba and, indirectly, in the near-extinction of the vaquita porpoise caught in nets set for it.
In 2014, under a binational agreement, a deliberate pulse flow was released to the delta — enough water to reconnect the river to the sea briefly for the first time in years. Vegetation responded measurably, and subsequent smaller releases have continued. It is restoration at a scale that acknowledges the delta cannot be brought back, only kept partly alive.
The rights that were counted last and rank first
American water law in the west runs on prior appropriation: the earliest claim has the strongest right, and in a shortage, junior users are cut before senior ones. That principle produces an awkward result on the Colorado.
Thirty federally recognised tribes hold rights in the basin, and under the doctrine established by the Supreme Court in Winters v. United States in 1908, those rights date from the establishment of their reservations — which in most cases predates the cities and irrigation districts that now depend on the river. Tribal rights are therefore among the most senior in the system, and collectively they amount to a substantial share of the river's flow.
For most of the twentieth century this was largely theoretical, because many tribes lacked the infrastructure to divert and use the water they were legally entitled to, and the 1922 Compact negotiations did not include them. Unused senior rights were absorbed in practice by junior users downstream. As tribes have quantified and begun to exercise those entitlements — through settlements, litigation and leasing arrangements — the arithmetic of the shortage has changed, and the negotiations over cuts now involve parties who were not at the table when the river was divided.
The accidental sea the river made
In 1905 the Colorado breached an irrigation headworks in the Imperial Valley and, for roughly eighteen months, the entire river poured into a dry basin below sea level in southern California. When engineers finally closed the breach, they had created the Salton Sea — California's largest lake, entirely by accident.
Having no outlet, it has survived since on agricultural runoff from the surrounding irrigation districts. That has two consequences: it is steadily getting saltier as water evaporates and salts concentrate, and it shrinks whenever less water is applied to the fields. Water transfers moving Imperial Valley allocations to coastal cities have accelerated the decline.
The exposed lakebed is the problem. It contains a century of accumulated agricultural residues, and wind lifts it as dust into one of the more asthma-burdened populations in California. A lake that exists only because of an engineering failure now requires continued water deliveries to keep its dust down — an obligation nobody planned for and everybody would prefer to hand to someone else.
The Murray, on the other side of the world
The closest parallel to the Colorado is not another American river but Australia's Murray, and the resemblance goes beyond both being dry-country rivers:
- Both were divided among jurisdictions using flow figures measured in unusually wet periods.
- Both have a downstream party — Mexico, South Australia — receiving what upstream users leave.
- Both routinely fail to reach the sea in meaningful volume, and both have had emergency interventions at the mouth.
- Both allocate most of their water to agriculture worth far less per litre than urban use, and both find that reallocating it is politically near-impossible.
The instructive difference is storage. The Colorado's reservoirs hold several years of average flow, which buffered the over-allocation for decades and disguised the problem; the Murray's hold much less, so its crises arrived sooner and more visibly. Large storage does not fix an over-allocated river — it postpones the reckoning and adds evaporative losses in the meantime.
What the river is now
The Colorado today is best understood as a plumbing system with a canyon attached. Its flow is scheduled rather than natural; its floods are deliberate; its temperature is set by the depth of a reservoir outlet; its sediment sits behind two dams; and its terminus is a negotiated allocation rather than an ocean.
None of that makes it less consequential. It waters Phoenix, Las Vegas, Los Angeles, San Diego, Denver and Salt Lake City, and irrigates a large share of American winter vegetables from the Imperial and Yuma valleys. The question occupying its basin is not whether the shortfall is real — that is settled — but which of those uses absorbs the reduction, and how much of the answer is decided by negotiation rather than by the reservoirs simply running out.
