For most of the past century, the Colorado River has ended in the sandy desert of northwestern Mexico, evaporating and soaking into the ground tens of kilometers before it reaches the Gulf of California. Historically, the river’s terminus was a vast, lush estuary where fresh water met the sea, teeming with fish and migratory birds. Today, reaching the ocean is the exception rather than the rule, and the stretches of dry riverbed below the U.S.-Mexico border tell the story of how thoroughly the river has been claimed by human demand.
How the River Lost Its Path to the Sea
The Colorado River begins high in the Rocky Mountains of Colorado and flows roughly 2,330 kilometers through the American Southwest before crossing into Mexico. Along the way, it passes through some of the driest terrain on the continent, and nearly every drop is spoken for. The river provides water to seven U.S. states and to Mexico under a system of treaties and agreements rooted in the Colorado River Compact of 1922, which divided the water based on flow measurements taken during an unusually wet period.1Proceedings of the National Academy of Sciences. The Legal Battle over the Colorado River Compact: Revisiting Water Allocation Agreements The allocations promised more water than the river typically carries, and the gap between what was promised and what actually flows has widened over time.
A series of massive dams and reservoirs built throughout the twentieth century fundamentally changed the river’s behavior. Hoover Dam, completed in 1935, created Lake Mead. Glen Canyon Dam, completed in 1963, created Lake Powell. Together with dozens of smaller dams and hundreds of diversion canals, these structures capture virtually the entire flow of the Colorado before it reaches Mexico. Cities like Las Vegas, Phoenix, Los Angeles, and San Diego depend on the river, as do millions of acres of irrigated farmland. Mexico is entitled to a share under a 1944 treaty, but by the time agricultural diversions in Mexico’s Mexicali Valley are finished drawing from the river, little or nothing remains to flow to the coast.
The physical consequences of damming started early. The upstream roughly 50 kilometers of the delta experienced deep channel incision beginning in the early 1940s, when spillway releases from Hoover Dam carried water but almost no sediment downstream.2Ecological Engineering. Geomorphic change and sediment transport during a small artificial flood in a transformed post-dam delta: The Colorado River delta, United States and Mexico Without sediment to build and replenish the floodplain, the channel cut deeper into its bed. The broad, braided river that once spread across a delta nearly two million acres wide became a narrow, entrenched channel surrounded by dry, salty ground.
What the Delta Looks Like Now
The Colorado River delta once supported dense galleries of cottonwood and willow trees, expansive wetlands, and a productive estuary where the river met the Gulf of California. The indigenous Cocopah people lived and fished there for centuries. By the mid-twentieth century, most of that ecosystem had collapsed. The cottonwoods died back as groundwater tables dropped. Soil salinity rose without periodic flooding to flush salts away. Into the gap moved saltcedar, an invasive shrub from Eurasia that tolerates high salinity and can outcompete native trees when river flows are reduced or eliminated.3Conservation Biology. Regeneration of Native Trees in the Presence of Invasive Saltcedar in the Colorado River Delta, Mexico
Saltcedar now dominates large sections of the riparian corridor. The connection between soil salinity and invasive takeover is well documented: as rivers are channelized and flooding eliminated, salt builds up in the soil to levels that native cottonwood and willow cannot handle, while saltcedar thrives.4Journal of Arid Environments. Growth rates, salt tolerance and water use characteristics of native and invasive riparian plants from the delta of the Colorado River, Mexico The result is a landscape that, to an untrained eye, may still look green in places but is ecologically impoverished compared to what existed before the dams.
Groundwater beneath the delta region has become increasingly important to both the U.S. and Mexico for meeting water demands, yet these shared aquifers rarely receive formal consideration in the official water-sharing agreements that govern the river.5Groundwater for Sustainable Development. Environmental evidence of surface manifestations of regional groundwater flows in the lower Colorado River Basin: the case Mexicali Valley, Mexico That means the below-ground water that sustains remaining patches of vegetation in the delta is being pumped without coordinated management, adding another layer of stress to an already depleted system.
When the River Does Reach the Sea
Despite the dams and diversions, the Colorado River has occasionally made it all the way to the Gulf of California in recent decades. After Lake Powell finally filled in 1981, unusually large inflows during El Niño years had to be released downstream because there was no remaining reservoir capacity to store them. Flood flows reached the delta channel during these wet cycles and brought a temporary revival of native vegetation along the riparian corridor.6Journal of Arid Environments. Ecology and conservation biology of the Colorado River Delta, Mexico The 1983 floods were particularly large, followed by significant flows in 1993 and again during the strong El Niño of 1997–98.
Direct salinity measurements at the mouth of the river during the 1993 and 1998 flood flows confirmed that fresh water actually reached the sea. Stable oxygen isotope signatures preserved in clam shells and fish otoliths recorded the dilution of seawater with Colorado River water during those events.7PubMed. Just add water and the Colorado River still reaches the sea These findings showed that the river’s plumbing still works when enough water enters the system; the problem is not that the channel is physically blocked but that the water is used up before it gets there.
These episodic floods also triggered regeneration of native cottonwood and willow trees along the delta corridor, even in areas dominated by saltcedar. The native trees germinated and grew in response to soil moisture delivered by flood flows, demonstrating that the delta retains the seed bank and biological capacity to recover if water is provided.3Conservation Biology. Regeneration of Native Trees in the Presence of Invasive Saltcedar in the Colorado River Delta, Mexico But these events are infrequent, and between floods, the gains are fragile.
What a Dry Estuary Has Done to Marine Life
The Gulf of California at the river’s mouth was once a productive estuary, with nutrient-rich fresh water mixing with seawater to create ideal nursery habitat for fish and shrimp. Losing that estuary has had consequences that researchers are still measuring. One of the most striking examples involves the totoaba, a large marine fish found only in the Gulf of California and now critically endangered. By comparing growth rings in otoliths (the tiny ear bones of fish) from prehistoric specimens dating back 1,000 to 5,000 years with those from modern fish, researchers found that pre-dam juvenile totoaba grew twice as fast and matured one to five years earlier than post-dam fish.8Biological Conservation. Diverting the Colorado River leads to a dramatic life history shift in an endangered marine fish Oxygen isotope analysis linked the change directly to the elimination of estuarine habitat. Without freshwater inflow, the juveniles lost the brackish-water nursery they evolved to depend on, grew more slowly, and took years longer to reach reproductive age.
The totoaba story is a window into broader ecological losses. The upper Gulf of California once supported a rich fishery partly sustained by Colorado River nutrients. The vaquita, the world’s most critically endangered marine mammal, lives in these same waters. While the vaquita’s decline is primarily driven by gillnet fishing, the overall degradation of the upper Gulf ecosystem through reduced nutrient input and changed salinity patterns has made the entire region less productive than it once was.
The 2014 Pulse Flow Experiment
In 2012, the United States and Mexico signed a binational agreement known as Minute 319, an amendment to the 1944 Water Treaty. Among other provisions, it authorized a deliberate release of water into the Colorado River delta to test whether ecological restoration was possible. The release happened in the spring of 2014: roughly 132 million cubic meters of water flowed into the delta corridor over eight weeks beginning in late March.9Ecological Engineering. A history of the 2014 Minute 319 environmental pulse flow as documented by field measurements and satellite imagery The flow was designed to mimic a spring flood, with the highest volumes released early to wet the channel and floodplain.
Most of the water, about 102 million cubic meters, was released at Morelos Dam on the U.S.-Mexico border. Additional water entered the corridor farther downstream through Mexican irrigation spillway canals.9Ecological Engineering. A history of the 2014 Minute 319 environmental pulse flow as documented by field measurements and satellite imagery The design itself was a novel collaboration involving federal and state water managers, local authorities, and conservation organizations from both countries who jointly shaped the flow delivery plan.10Ecological Engineering. Shaping the 2014 Colorado River Delta pulse flow: Rapid environmental flow design for ecological outcomes and scientific learning
The pulse flow was tiny by historical standards. The Colorado River’s annual natural flow before damming was estimated at around 18 billion cubic meters. The 2014 release amounted to less than one percent of that. But the experiment demonstrated that even a small volume of water could produce measurable ecological responses. The river flowed through previously dry stretches of the delta, wetted soils and recharged shallow groundwater, and brought a visible green-up visible in satellite imagery.11Ecological Engineering. Leveraging environmental flows to reform water management policy: Lessons learned from the 2014 Colorado River Delta pulse flow
Why the Pulse Flow’s Green-Up Was Hard to Sustain
The ecological results of the 2014 pulse flow were more complicated than the satellite photos suggested. Without active land management, such as clearing competing vegetation and preparing bare ground for germination, the flows were only minimally successful at establishing new native woody seedlings. In many areas, the primary beneficiary of the added water was saltcedar rather than cottonwood or willow.12Ecological Engineering. A large-scale environmental flow experiment for riparian restoration in the Colorado River Delta The reasons varied by location but included a lack of native seed availability, groundwater tables that dropped too quickly after the pulse ended, and competition from existing vegetation that crowded out seedlings.
Where restoration teams had cleared invasive plants and prepared the ground beforehand, cottonwood and willow did establish, though results were patchy. Successful germination depended on a cluster of factors: wetted soils, bare surface availability, low soil salinity, and seeds arriving at the right time. Even then, seedling survival over the growing season depended on soil conditions remaining favorable. Saltcedar germinated and persisted across a much wider range of conditions, including in areas that had not been cleared.13Ecological Engineering. Integrating active restoration with environmental flows to improve native riparian tree establishment in the Colorado River Delta The relatively small volume of water released could not restore natural flooding processes like erosion and sediment deposition that historically maintained the delta’s ecosystem.
The honest takeaway was that water alone is not enough. Restoring the delta requires pairing water deliveries with on-the-ground restoration work, and even then, the invasive species advantage is steep. The pulse flow proved the concept that ecological restoration in the delta is possible, but it also made clear that the scale of effort needed to produce lasting change is far greater than a single eight-week release.
A Shrinking River in a Warming Climate
The Colorado River’s supply problem is getting worse. The river’s flow is sensitive to temperature: warmer air pulls more moisture from snowpack and soil before it can become runoff. Research has found that the Colorado Basin’s runoff has decreased by roughly 8 percent for every degree Celsius of warming.14Water Resources Research. Aridification of Colorado River Basin’s Snowpack Regions Has Driven Water Losses Despite Ameliorating Effects of Vegetation Since 1880, anthropogenic warming has reduced the basin’s natural flow by an amount roughly equivalent to the storage capacity of Lake Mead. The 2000–2021 megadrought, the worst in over a thousand years, was made dramatically worse by this warming trend. Researchers have concluded that the basin’s first official shortage declaration in 2021 likely would not have occurred without human-caused climate change.14Water Resources Research. Aridification of Colorado River Basin’s Snowpack Regions Has Driven Water Losses Despite Ameliorating Effects of Vegetation
The losses are concentrated in the places that matter most. Snowpack regions make up only about 30 percent of the basin’s drainage area but account for roughly 86 percent of the runoff decreases. Snow at high elevations melts and evaporates faster in warmer conditions, and the runoff decline in these snowpack zones is double the rate seen in lower, non-snowpack regions. Looking ahead, projections are sobering: if warming continues on its current trajectory, temperature-driven flow declines could conservatively reach 20 percent by mid-century and 35 percent by the end of the century, with some projections supporting losses exceeding 50 percent.15Water Resources Research. The twenty‐first century Colorado River hot drought and implications for the future
Evaporation from the reservoirs themselves compounds the problem. Lake Mead alone loses an average of roughly 1,900 millimeters of water per year to evaporation, about six feet off its surface annually.16U.S. Geological Survey. Evaporation from Lake Mead and Lake Mohave, Lower Colorado River Basin, Nevada and Arizona As the reservoirs shrink in area, the absolute evaporative loss decreases somewhat, but it remains a significant drain on a system where every drop is contested.
Lake Mead, Dead Pool, and What Comes After 2026
The current operating guidelines for the Colorado River expire after 2026, and negotiations over what comes next are underway among the seven basin states, tribal nations, and Mexico. The stakes are enormous. Modeling of different proposed policy frameworks shows that even the most aggressive water-conservation approach leaves a 54 percent probability that Lake Mead hits “dead pool” at least once between 2060 and 2100, meaning the water level drops below the lowest outlet and the reservoir can no longer release water downstream.17Nature Communications. Disentangling climate and policy uncertainties for the Colorado River post-2026 operations Under a scenario closer to existing rules, the cumulative risk climbs above 90 percent. In any given year before 2060, the probability of dead pool under the best-case policy is around 5 percent, but risks compound over time.
These projections mean that the question of where the Colorado River ends is likely to become even more stark. If Lake Mead reaches dead pool, the river’s managed flow to southern Nevada, Arizona, and the Mexican border effectively stops. The river would still flow below Glen Canyon Dam when Lake Powell has water to release, but the lower river system and everything downstream of it, including the delta, would face an unprecedented disruption. Researchers have emphasized that governance frameworks need to be adapted to account for climate variability, reduce overall consumption, and strengthen participation by tribal governments whose water rights have historically been underrepresented in negotiations.18BioScience. Reimagining river governance: Insights from the Colorado river crisis
Sediment, the Forgotten Loss
When people talk about the Colorado River running dry, they focus on water. But the river also used to carry enormous quantities of sediment, the fine red silt that gave the river its Spanish name (colorado means “colored” or “ruddy”). The dams trap virtually all of that sediment in their reservoirs. Lake Powell alone has accumulated billions of tons of silt since Glen Canyon Dam was completed. Downstream, the river runs relatively clear, which sounds pleasant but is ecologically destructive. Without sediment, the river cannot build sandbars, replenish beaches, or maintain the kind of shifting floodplain habitat that cottonwoods need to colonize. The channel incision that began in the 1940s below Hoover Dam was a direct consequence of this sediment starvation, as clear water released from the dam scoured the channel deeper rather than spreading across the floodplain.2Ecological Engineering. Geomorphic change and sediment transport during a small artificial flood in a transformed post-dam delta: The Colorado River delta, United States and Mexico
The 2014 pulse flow illustrated this limitation vividly. Even though water traveled through the delta corridor, the flow was too small to move significant amounts of sediment or reshape the channel. The natural processes of erosion and deposition that historically maintained the delta’s mosaic of habitats simply did not happen at the scale of the experimental release.13Ecological Engineering. Integrating active restoration with environmental flows to improve native riparian tree establishment in the Colorado River Delta Restoring water flow is necessary but insufficient if the sediment that shapes habitat is permanently locked behind dams hundreds of kilometers upstream. This is a constraint that no amount of policy negotiation over water volumes can solve on its own, and it makes the task of bringing the delta back to anything resembling its historical condition far more daunting than simply finding water to release.