Why in news?
A study estimates larger groundwater losses in the Upper Colorado River Basin than in the Lower Basin during the 2020–2022 drought. It was published in Water Resources Research on 13 October 2026. That is significant because this mountain-fed region usually receives more natural replenishment than the drier Lower Basin. The Colorado supplies farms, communities and ecosystems across the southwestern United States and northern Mexico, through a connected river and reservoir system. Groundwater, held beneath the surface in water-bearing material, can help sustain supplies when rivers run low. The finding suggests that this buffer also needs protection in the Upper Basin. It is a satellite-assisted estimate with uncertainty, not a measurement of every well. It does not predict that the entire river will dry on a particular date.
The river's course is not the same as its whole basin
The Colorado rises in the Rocky Mountains of north-central Colorado. It travels southwest through the Colorado Plateau, southeastern Utah and the canyon country of northern Arizona. Farther downstream it forms stretches of the Nevada–Arizona and Arizona–California boundaries before continuing into Mexico. Its historical outlet is the Gulf of California, between the Baja California peninsula and mainland Mexico. The Green, Gunnison, San Juan, Virgin, Little Colorado and Gila are important tributaries. These connections bring water from a much larger landscape than the narrow strip occupied by the main channel.
The seven basin states are Arizona, California, Colorado, Nevada, New Mexico, Utah and Wyoming. This does not mean that the main river flows through all seven. A basin is the land whose drainage contributes to a river system; a tributary can connect a state to that system. Reclamation's geographical division places the Upper Basin above Lees Ferry in Arizona and the Lower Basin below it. These physical catchments do not follow whole-state boundaries neatly. For example, part of Arizona drains to the Upper Basin, while parts of Utah and New Mexico drain to the Lower Basin.
Visible reservoirs reveal only part of the water system
Lake Powell, impounded by Glen Canyon Dam, is upstream of Lake Mead, held by Hoover Dam. The river also cuts through the Grand Canyon between the dam systems. Reservoirs make water storage visible, and their operations influence supplies, electricity generation and river conditions. They are, however, artificial surface stores within a larger water cycle. Knowing the amount of water in a reservoir does not establish how much remains in the surrounding ground. The geographical map identifies these features; it does not show their present water levels or the study's groundwater-loss pattern.
What groundwater is and how it reaches a river
Groundwater is not normally a vast open underground lake. It occupies connected pores and fractures in soil and rock. A formation that stores and transmits groundwater is called an aquifer. Its saturated part contains water filling those spaces. The water table marks the top of that saturated zone where it is not confined beneath an impermeable layer. Rain and melting snow can infiltrate the ground and contribute to recharge, replenishing groundwater. Some precipitation instead runs over the surface or returns to the atmosphere. Consequently, precipitation, recharge and water available for use are different quantities.
A river can gain water where groundwater levels beside it are higher than the stream's water surface. This contribution, called baseflow, can keep a channel flowing between rainfall or snowmelt events. Elsewhere, water can move from a stream into the ground; the same river may gain and lose water in different reaches or seasons. This is why surface water and groundwater cannot always be managed as independent reserves. Drawing down one part of a connected system can affect another, even when the connection is not visible from the bank.
What the new drought comparison actually measures
Esther Oyedele and colleagues estimate groundwater-storage losses of 17.5 ± 3.5 cubic kilometres in the Upper Basin. Their Lower Basin estimate is 9.4 ± 6.5 cubic kilometres. These are the study's estimates for water years 2020–2022. A United States water year runs from 1 October through 30 September and is named for its ending year. The comparison therefore covers October 2019 to September 2022, not three calendar years beginning in January 2020. The central estimates differ, but the stated uncertainty should not disappear behind a precise “twice as much” headline.
A cubic kilometre is a volume: one billion cubic metres. It is not a fall of one kilometre in the water table. Water-level changes measured at wells and storage changes across a basin describe related but different things. Translating between them requires information about the area, depth and water-holding properties of the aquifer. Similarly, a smaller estimated loss does not automatically identify a healthier reserve. A region starting with less recharge or a long history of depletion can remain vulnerable even if its change during one drought is smaller.
Why natural replenishment does not guarantee resilience
Virginia Tech's account highlights the Upper Basin's usual snowmelt-driven recharge and its stronger connection between surface-water shortages and groundwater decline. During a prolonged dry period, groundwater may continue feeding streams while its replenishment weakens. The store can therefore help maintain river flow while being depleted itself. The researchers contrast this with the drier Lower Basin, where sustained human water use can exert pressure. That pressure can be less closely tied to immediate surface-water changes. This is a difference in the pattern of vulnerability, not evidence that either half can disregard groundwater demand.
Resilience depends on recovery as well as the ability to supply water during a dry spell. If a reserve is repeatedly used without sufficient replenishment between droughts, its buffering role becomes harder to sustain. A wet season can improve some conditions without demonstrating that every aquifer has recovered. River flow, snow storage, soil moisture and groundwater also respond on different timescales. These distinctions help explain why a favourable surface-water snapshot cannot settle the longer-term water-security question.
How satellites help estimate water beneath the surface
The Gravity Recovery and Climate Experiment, or GRACE, and its Follow-On mission provide observations used in the study. Their technique is not an underground camera. Water has mass, and changes in the distribution of mass affect Earth's gravity. Precise measurements involving the satellite pair allow scientists to estimate changes in stored water across large areas. The signal includes different stores, such as snow, soil moisture, reservoirs and groundwater. An observed gravity-related change must therefore be interpreted with additional information before it is described specifically as groundwater loss.
Scientists combine these observations with other measurements and hydrological models to separate the components. Estimating the snow, surface-water and soil-moisture contributions allows groundwater change to be inferred from the remaining storage signal. Uncertainty in those components also affects the groundwater estimate. The approach provides a broad view that individual wells alone cannot supply, but it does not replace local measurements. Nor does a regional storage pattern precisely divide each household's or farm's shortage into pumping and climate effects.
Managing the connected system
The practical implication is to assess the water above and below ground together. Reservoir operations address visible stores and releases; groundwater monitoring helps reveal whether the less visible reserve is being replenished. Local well records, stream measurements and regional satellite estimates answer different questions and can be used together. Demand also needs to be considered alongside replenishment. Moving a shortage from river withdrawals to wells may change where water is taken without resolving the underlying imbalance in a connected system.
Conclusion
The study brings attention to a less visible part of Colorado River water security: groundwater in the supposedly more resilient mountain-fed Upper Basin. The important finding is not that the Lower Basin is safe, or that every local aquifer behaves alike. It is that a drought buffer can be depleted while it helps sustain streams and supplies. Geographical context, clear storage measurements and uncertainty-aware monitoring are needed to judge that risk. A fuller reservoir alone cannot demonstrate that the whole river–groundwater system has recovered.