On 24 November 1922, the Colorado River Commission officially allocated water rights to the seven U.S. states of the Colorado River Basin. The Colorado River Compact and subsequent agreements, collectively known as the Law of the River, eased years of dispute among these states, and they constitute a milestone in the history of the American West.
The 1922 compact provided regulatory certainty for water management. It called for water to be stored and released as needed (most notably with the construction of Hoover Dam), thus supporting a robust era of reservoir building. The reservoirs, in turn, unleashed huge potential for electric power generation and stimulated economic growth throughout the West.
The terms of the compact, however, were largely the product of development aspirations and political dealmaking, and they relied on optimistic estimations of the amount of water the river could supply that were not supported by existing surveys or science. One hundred years later, a lasting water shortage crisis has brought the governance structure outlined in the compact to its knees, and the effects reverberate far beyond the Colorado River Basin. The two largest reservoirs in the United States, Lake Mead and Lake Powell, have reached historic lows, threatening both the water supply and the hydropower generation capacity for tens of millions of users, as well as the nation’s food supply and flows critical to maintaining ecosystems.
Municipalities are considering drastic water-saving measures. Farmers and ranchers, who as a group are by far the largest consumers of Colorado River water, face unprecedented challenges and uncertainty. So do the economies and environmental systems that depend on reliable stream discharge for aquatic life and recreation.
It is tempting to place responsibility for the water shortages on climate change, which has resulted in reduced precipitation across the basin, and on population growth that outpaced planners’ anticipation of water demands. Indeed, these are important exacerbating factors. A root cause of the dire situation today, however, lies in the commission’s choice to ignore the best available hydrologic science as it negotiated the original compact. As discussions over the availability of Colorado River water continue and a new compact is negotiated over the next few years, planners must not make this mistake again.
The Law of the River
Even before the Colorado River Compact was established, the vast American West was a bustling frontier for mineral exploration, agricultural development, and westward expansion. California already had been diverting water from the Colorado River to irrigate the fertile Imperial Valley since around 1901. Agriculture in sunny but dry southern Arizona was also booming. Other states envisioned securing more water for future irrigation of farmlands and for urban development.
Members of the commission included eight men, one each representing the Colorado River Basin states—Arizona, California, Colorado, Nevada, New Mexico, Utah, and Wyoming—plus the commission’s chair, Secretary of Commerce Herbert Hoover, who later became president of the United States. All parties realized the paramount importance of agreeing on consistent apportionments of the river’s water to the states, which would provide needed certainty into the future [Kuhn and Fleck, 2019]. That meant estimating the magnitude of the river’s discharge.
The main elements of the compact included the following:
- The Colorado River Basin was divided into the Upper and Lower basins at Lee Ferry, Ariz. The Upper Basin includes four states: Colorado, New Mexico, Utah, and Wyoming. The Lower Basin includes three: Arizona, California, and Nevada (Figure 1).
- Consumptive water use was divided evenly between the Upper and Lower basins, with each allowed 7.5 million acre-feet (~9.2 billion cubic meters) per year. The Upper Basin states were obligated to “not cause the flow of the river at Lee Ferry to be depleted below an aggregate of 75,000,000 acre-feet for any period of ten consecutive years.”
- The river’s average discharge at Lee Ferry was assumed to be 16.4 million acre-feet per year. Allocating a total of 15 million acre-feet per year would leave the remaining water for future development and for Mexico.
The allocation of 7.5 million acre-feet per year of consumptive use for each basin was grounded neither in the best available hydrologic calculations nor in climate variability projections. Rather, it was a compromise Hoover proposed between two endmember figures [Kuhn and Fleck, 2019]. One end was 8.2 million acre-feet per year, half the assumed annual average discharge at Lee Ferry of 16.4 million acre-feet per year, which itself was derived from a report by the U.S. Reclamation Service (now the Bureau of Reclamation) [Fall and Davis, 1922]. The other end was 6.5 million acre-feet per year, a figure proposed and advocated by the Upper Basin commissioners that reflected a roughly 50-50 split of the river discharge at Yuma [Kuhn and Fleck, 2019].
In the decades following the 1922 compact, a plethora of acts, orders, and agreements were written and signed to fine-tune the compact’s provisions, to authorize construction of dams for water storage and power generation, to build water transfer infrastructures, and to resolve interstate and intrastate disputes. Particularly significant was the 1944 treaty between the United States and Mexico—still in effect today—that guaranteed 1.5 million acre-feet of Colorado River water annually for Mexico, bringing the total allocation to 16.5 million acre-feet per year.
How Much Water Was There?
In the early 1900s, there were only a few stream gauges in the United States measuring river discharge. The middle section of the Colorado River Basin was one of the most remote and inaccessible regions in the nation at the time. In particular, the canyon region from the mouth of Green River in Utah to the Grand Wash in Arizona, covering a water course of approximately 840 kilometers (520 miles), was accessible to wheeled vehicles at only three points [La Rue et al., 1925]. Because of the inaccessibility, no stream gauges were established there until about 1920. The gauge station at Lee Ferry was established only in summer 1921.
The estimate of Colorado River discharge at Lee Ferry adopted in the compact originated with stream discharge measurements at Laguna Diversion Dam (Figure 2) near Yuma in southern Arizona, a water course of approximately 1,002 kilometers (622 miles) downstream of Lee Ferry. Fall and Davis  derived the value by subtracting the discharge from the Gila River, which enters the Colorado at Yuma, from the measured discharge at Laguna Dam. The commission simply assumed that the volume gained by the Colorado from tributaries between Lee Ferry and Laguna Dam was about the same as the volume lost to evaporation over that same stretch of river corridor (black curve in Figure 2).
Even today, it is challenging to estimate water loss due to evaporation and plant transpiration over a vast area of dry land influenced by seasonal floods and varying vegetation covers. It’s clear that the commission’s assumption, and therefore the 16.4-million-acre-feet-per-year estimate, was informed by grossly optimistic considerations and ignored the more conservative science and more reliable hydrologic data available at the time.
A lower estimate of Colorado River discharge had emerged prior to the compact on the basis of a more rigorous scientific approach by U.S. Geological Survey (USGS) hydrologist Eugene Clyde La Rue [Kuhn and Fleck, 2019]. Between 1914 and 1924, La Rue traveled hundreds of miles of the Colorado River and its tributaries to survey dam sites and conduct river discharge measurements. He probably collected more firsthand hydrologic data than anyone and was considered the most knowledgeable Colorado River expert of his generation [Langbein, 1975].
La Rue calculated the average discharge at Lee Ferry between 1895 and 1920 to be 15.0 million acre-feet per year using records from stream gauges and tributary contributions upstream of Lee Ferry. Specifically, he used upstream gauges near Green River, Utah, on the Green River and near Fruita, Colo., on the Colorado River (Figure 1), combined with his records from several other tributaries, to estimate the discharge at Lee Ferry (red curve in Figure 2).
How significant is this difference? It represents nearly 10% of the river discharge assumed in 1922, and it is not far below the estimated reduction in demand needed to meet the current shortage. At a U.S. Senate committee hearing in 2022 examining short- and long-term solutions to extreme droughts in the western United States, Bureau of Reclamation commissioner Camille C. Touton testified, on the basis of a bureau analysis, that Colorado River Basin states would need to reduce consumption by 2–4 million acre-feet in 2023 to protect hydropower generation at Lake Mead and Lake Powell.
La Rue argued that decisionmakers should use longer-term averages in estimating river discharge. Prior to 1899, there were no stream discharge measurements in the Colorado River Basin. La Rue creatively used water level records from Great Salt Lake in Utah, calibrated against later records of river discharge and lake levels, to infer earlier annual Colorado River discharges back to 1895 [La Rue and Grover, 1916; La Rue et al., 1925] (dashed red curve in Figure 2). Decades later, La Rue’s inferred discharges for those early years were found to be consistent with discharge values estimated from tree ring studies [Meko et al., 2007]. The different approaches of La Rue and Fall and Davis led to a disparity in their discharge estimates of approximately 1.4 million acre-feet per year.
Ignoring Available Science
Data and science characterizing the Colorado were limited in the 1920s, but La Rue’s river discharge estimate was known ever since he first published it in a USGS report in 1916. Yet his work only hovered in the background of the commission’s negotiations. La Rue made a series of attempts to let the commission know that its perception of how much water was in the river was overly optimistic [Kuhn and Fleck, 2019]. In 1920, he tried but failed to facilitate a meeting between USGS and the Reclamation Service because of his concerns over the difference between his estimate, published in the 1916 USGS report, and Fall and Davis’s estimate, which first appeared in a preliminary Reclamation Service report in 1920.
As the preparation of the compact was gathering steam, La Rue took the unusual step of writing directly to Secretary Hoover. He received only a thank-you note in return. The commission refused to be distracted by any lower estimate of river discharge and forged ahead with the compact based on Fall and Davis’s estimate. The higher estimate, of course, meant more perceived water for everyone, which understandably would make negotiations easier. Whether the commission fully recognized the potential consequences of its inattention to and dismissal of La Rue’s lower estimate at the time is unclear.
In 2007, the Colorado River Basin was experiencing the worst 8-year period of drought in more than a century of continuous recordkeeping. The U.S. Department of the Interior (DOI) issued interim guidelines to address issues related to Lower Basin water shortages and the management of the Lake Mead and Lake Powell reservoirs. These guidelines encouraged voluntary water conservation measures but did not attempt to reallocate water deliveries to compact states.
More than a decade later, as the drought continued, the combined storage in Lake Powell and Lake Mead reached its lowest volume since the 1960s. In 2019, the Bureau of Reclamation then established a Drought Contingency Plan, setting an example of collaboration across the Colorado River Basin. The plan required Upper and Lower basin states to work together to address the imminent water crisis and better manage the Colorado River system in the future.
The Lower Basin approach in the Drought Contingency Plan focused on reducing water demand to stabilize water levels in Lake Mead, while the Upper Basin approach focused similarly on maintaining water levels in Lake Powell. The plan also offered recommendations for voluntary water conservation programs to compensate farmers and other water users for reducing their water use without losing their water rights under the Prior Appropriation doctrine.
In May 2023, DOI announced a deal agreed upon by the three Lower Basin states to conserve at least 3 million acre-feet of water through 2026 to maintain reservoirs above critical levels. Of that amount, 2.3 million acre-feet will be compensated through funding from the Inflation Reduction Act to support water conservation efforts and enhancements to water system efficiency. The remaining conservation needed for sustainable operation will come from voluntary and uncompensated reductions by the Lower Basin states.
Bring Science—and All Parties—to the Table
The interim management guidelines established in 2007 are set to expire in 2026, the date set for review and reauthorization of the 1922 compact. Between now and 2026, there is a window of opportunity to rebalance the allocation and availability of water. It is time to confront the fact that the combination of natural flow and reservoir storage on the Colorado does not provide enough water to meet current demands, as La Rue recognized 100 years ago. It is time to bring science to the negotiating table.
There is no shortage of examples where science has successfully informed water management policy [Loucks, 2021]. Consider the collaboration between Canada and the United States to manage Lake Ontario and the St. Lawrence River. In response to changing needs of various sectors (e.g., recreation, commercial fishing) and natural hydrologic conditions, the International Lake Ontario–St. Lawrence River Board  conducted a comprehensive multiyear study to guide revisions to the existing 50-year-old regulations on water levels and river flows for hydropower generation, river navigation, and flood controls.
The board engaged the public and experts, addressed issues pertinent to affected Indigenous communities as an integral part of the process, and applied state-of-the-art scientific knowledge to inform the discourse over new regulations.
For example, the 2006 study found that shoreline communities preferred lower lake levels, which minimize damages from flooding and erosion, whereas recreational users preferred higher levels. Meanwhile, scientific research considered in the study indicated that widely varying lake levels in the Great Lakes are favorable for healthier ecosystems [e.g., Wilcox et al., 2007]. Together the findings required the board to rethink the interests of shoreline communities and recreational users and of how to maintain reliable water intakes for hydropower. The board then devised regulation options that would benefit a greater number of interest groups than the current regulations did and minimize losses for any single group or geographical area. The study also developed adaptation alternatives to help manage climate change–driven uncertainties in future conditions.
There is no doubt that climate change, droughts, and population growth have exacerbated the Colorado River water shortage crisis. It is also obvious that the best available science was ignored 100 years ago and water from the Colorado River was overallocated. Negotiators today must learn from history and embrace state-of-the-art science to help reallocate the Colorado River sustainably.
Long-term up-to-date natural discharge data at Lee Ferry are available (Figure 2). As of 2022, the 20-year running average stands below 13 million acre-feet per year, a 20% reduction from what was assumed in the original compact. Further decreases are expected.
Milly and Dunne , considering a moderate greenhouse gas emission scenario (i.e., Representative Concentration Pathway 4.5), predicted that average discharge from the Upper Colorado River Basin between 2016 and 2065 could be 5%–24% less than it was in 1903–2017. Miller et al.  projected a 5% decrease at Lee Ferry for the period 2040–2069 relative to 1975–2005. Li and Quiring  projected that discharge in the Upper Colorado River Basin will decrease 2.3%–21.0% due to climate and land use change from 2040 to 2069. The 16.4-million-acre-feet-per-year figure, an overestimate in 1922, is far from realistic today and in the foreseeable future.
In addition to considering the best available science, all stakeholders—notably including those left out of the 1922 compact—must have seats at the negotiating table. Twenty-nine Native American tribes in the Colorado River Basin were granted rights to water for their reservations by the United States Supreme Court in Winters v. United States (1908). And the 1922 compact states: “Nothing in this compact shall be construed as affecting the obligations of the United States of America to Indian tribes.” Yet the agreement made no explicit allocations because tribal representatives were not present during the negotiations, and no subsequent water deliveries were made because there was no infrastructure to convey water to tribal lands.
Formally incorporating tribal water rights is a necessity in the challenge of reallocating the Colorado River. A 2018 study conducted jointly by the Colorado River Basin Ten Tribes Partnership and the Bureau of Reclamation found that the 29 tribes have enforceable rights to as much as 2.8 million acre-feet of Colorado River water per year [U.S. Bureau of Reclamation, 2018], or more than 20% of the 13-million-acre-feet-per-year recent average flow.
Similarly, the treaty rights of Mexico to Colorado River water also must be included. These rights are mandated by a standing international agreement and are a model for needed bilateral collaboration on allocating the water of the Rio Grande River, which has tributary headwaters in Mexico and the United States.
Reducing long-term regional allocations will be unpopular, but it is a necessity that negotiators need to accept. The reduced allocations must be embedded in the new compact, and whether as percentages of the natural discharge or of specific volumes, they must be based on robust estimates grounded in the best hydrologic and climate science available.
The total allocation also must account for all stakeholders and reflect expected declines in discharge over the coming decades. Furthermore, decisions and agreements on reallocation should precede regional- and local-scale actions taken to reduce water use, such as conservation, land use changes, water reuse, and water transfers, so that these actions can be implemented according to revised allocations.
Existing tools used to confront the water crisis, which have been used mostly on a volunteer basis and/or on local scales, have achieved limited success, attesting to the difficulty of the choices ahead and to the need for broader, more enforceable regulations. Remembering E. C. La Rue’s science-based approach and thinking long term will bring much to current negotiations and help sensibly reenvision the Colorado River Compact.
Fall, A. B., and A. P. Davis (1922), Problems of Imperial Valley and vicinity, 326 pp., U.S. Gov. Print. Off., Washington, D.C., https://hdl.handle.net/2027/hvd.32044031907595.
International Lake Ontario–St. Lawrence River Board (2006), Options for managing Lake Ontario and St. Lawrence River water levels and flows: Final report, 146 pp., Buffalo, N.Y., https://www.ijc.org/en/glam/options-managing-lake-ontario-and-st-lawrence-river-water-levels-and-flows-final-report.
Kuhn, E., and J. Fleck (2019), Science Be Dammed: How Ignoring Inconvenient Science Drained the Colorado River, 288 pp., Univ. of Ariz. Press, Tucson.
Langbein, W. B. (1975), L’Affaire LaRue, U.S. Geol. Surv. Water Resour. Div. Bull., April–June, 6–14.
La Rue, E. C., and N. C. Grover (1916), Colorado River and its utilization, U.S. Geol. Surv. Water Supply Pap., 395, 231 pp., https://doi.org/10.3133/wsp395.
La Rue, E. C., H. Work, and N. C. Grover (1925), Water power and flood control of Colorado River below Green River, Utah, U.S. Geol. Surv. Water Supply Pap., 556, 176 pp., https://doi.org/10.3133/wsp556.
Li, Z., and S. M. Quiring (2022), Projection of streamflow change using a time-varying Budyko framework in the contiguous United States, Water Resour. Res., 58(10), e2022WR033016, https://doi.org/10.1029/2022WR033016.
Loucks, D. P. (2021), Science informed policies for managing water, Hydrology, 8(2), 66, https://doi.org/10.3390/hydrology8020066.
Meko, D. M., et al. (2007), Medieval drought in the upper Colorado River Basin, Geophys. Res. Lett., 34(10), L10705, https://doi.org/10.1029/2007GL029988.
Miller, O. L., et al. (2021), Changing climate drives future streamflow declines and challenges in meeting water demand across the southwestern United States, J. Hydrol. X, 11, 100074, https://doi.org/10.1016/j.hydroa.2021.100074.
Milly, P. C. D., and K. A. Dunne (2020), Colorado River flow dwindles as warming-driven loss of reflective snow energizes evaporation, Science, 367(6483), 1,252–1,255, https://doi.org/10.1126/science.aay9187.
U.S. Bureau of Reclamation (2018), Colorado River Basin Ten Tribes Partnership tribal water study: Study report, U.S. Dep. of the Interior, Washington, D.C., https://www.usbr.gov/lc/region/programs/crbstudy/tws/finalreport.html.
Wilcox, D. A., et al. (2007), Lake-level variability and water availability in the Great Lakes, U.S. Geol. Surv. Circ., 1311, 25 pp., https://doi.org/10.3133/cir1311.
Shemin Ge (email@example.com) and Joann Silverstein, University of Colorado Boulder; James Eklund, Sherman & Howard LLC, Denver; Patricia Limerick, University of Colorado Boulder; and David Stewart, Stewart Environmental Consulting Group, Fort Collins, Colo.