Geologists can’t believe it: a US river appears to flow uphill – and they think they’ve found the reason

For well over a century, walkers, rafters and researchers have stood beside Utah’s Green River and wondered the same thing: why does it appear to run the “wrong” way through the Uinta Mountains, rather than taking the simpler option of going around them?

The Green River that seems to defy gravity

The Green River is anything but a minor watercourse. As one of the principal tributaries of the Colorado River, it feeds the wider system that helped carve the Grand Canyon and delivers water to tens of millions of people across the western United States.

Yet one section looks baffling on a map. Instead of skirting high ground, the river slices straight across the Uinta Mountains-a tough east–west range rising to more than 4,000 metres in north-eastern Utah and north-western Colorado. From certain viewpoints, the flow can even seem to be travelling “uphill” as it threads through a barrier of rock.

The Green River didn’t edge around the Uinta Mountains as most rivers do when faced with high terrain – it cut directly through them.

That apparent contradiction has fuelled debate among geologists for more than 150 years. The heart of the puzzle is timing: the Uinta range is about 50 million years old, but the river has excavated Lodore Canyon, a gorge roughly 700 metres deep, right through the range’s interior.

Classic geomorphology would predict a different outcome. Rivers generally adopt the easiest route, bending away from rising mountains and fault blocks. The Green River appears to have ignored that rule-so the obvious question has been why.

New research on the Green River points underground

A recent paper in the Journal of Geophysical Research: Earth Surface, led by Dr Adam Smith (University of Glasgow), argues that the river’s odd path is driven less by surface erosion alone and more by processes unfolding tens of kilometres beneath our feet.

The international group-bringing together experts from University College London and several US institutions-combined multiple lines of evidence:

• Seismic tomography, comparable to a medical CT scan but used to image Earth’s interior
  
• Numerical models that reproduce how rock deforms and flows over millions of years
  
• Detailed mapping and analysis of regional river networks

Their conclusion points to a deep geodynamic mechanism called lithospheric drip-a term unfamiliar outside geology, but one capable of subtly (and decisively) reshaping continental surfaces.

A practical implication is that river routes can be influenced by forces that never break the surface. That matters not only for interpreting landscapes, but also for anticipating where erosion concentrates, how sediment is delivered downstream, and why certain canyons form where they do.

The “uphill” effect is an illusion created by ancient sagging and later uplift of Earth’s crust beneath the river.

Related reads (as listed in the original)

• New spacecraft images reveal interstellar comet 3I ATLAS in far more detail than scientists expected
  
• China planted so many trees in the Taklamakan Desert that it now absorbs CO₂
  
• Japan crosses a strategic red line with a new stealth missile capable of mid-air corkscrew manoeuvres to evade defences and strike targets beyond 1,000 km
  
• Behavioural scientists say people who walk faster than average tend to be more successful and mentally sharper than slow walkers
  
• A psychologist insists the best stage of life begins when you start thinking in a particular way
  
• A new inheritance law coming into force in March is set to reshape key rules for heirs and families
  
• Unseen for 175 years: the world’s climate is tipping dangerously
  
• The Summer I Turned Pretty: the haircut to ask your hairdresser for to get Belly’s look
  

What lithospheric drip actually is

Earth’s rigid outer shell-the lithosphere-includes the crust and the uppermost mantle. In broad terms it “floats” on the hotter, more ductile mantle below.

Occasionally, part of the lower lithosphere becomes unusually dense. Over time, that heavier material can peel away and sink into the mantle-rather like a blob of cooler syrup dropping through warmer honey. That downward detachment is what geologists refer to as a lithospheric drip.

How a sinking rock mass steers a river

The study proposes that a lithospheric drip developed beneath the northern flank of the Uinta Mountains roughly 2–5 million years ago. In geological terms that is recent, particularly for a mountain range whose origins go back tens of millions of years.

As the dense material started to sink, the land surface above it drooped slightly. The result was a broad, gentle depression-not a dramatic crater, but a subtle sag sufficient to change which direction offered rivers the lowest route.

A temporary sag in the crust opened a low pathway across the mountains, and the Green River took it.

In other words, water follows the local gradient, even when that gradient runs counter to a wider regional tilt. The Green River adjusted to this newly created low corridor, and as it flowed it began to cut down into rock.

Once the dense patch had fully detached and sunk deeper, the crust began to lift again through isostatic rebound. By then, the river had already entrenched itself, carving a sizeable canyon and effectively fixing its course through the mountain belt.

Today the physics is ordinary-the river still runs downhill-but the route looks counterintuitive when compared with the older surrounding topography, which is why it can appear as if it is travelling “the wrong way”.

An important nuance is that this is not a story of a river “winning” a battle against mountains in a single episode. It is a sequence: deep sinking triggers sag, sag encourages a new pathway, incision locks that pathway in place, and rebound then raises the terrain around an already-established channel.

The making of Lodore Canyon

This deep-Earth sequence helps explain why Lodore Canyon exists at all. Now a well-known destination for rafters in Dinosaur National Monument, the canyon is presented by the study as a surface response to changes far below the crust.

The authors suggest that much of the canyon cutting happened in just the last few million years-fast in geological terms, especially given the ~50 million-year history of the range.

Using seismic imaging, the team also reports a deep seismic anomaly beneath the Uinta Mountains. They interpret it as a likely remnant of the dense material involved in the lithospheric drip.

The deep anomaly beneath the Uintas is a smoking gun linking mantle dynamics to the shape of the Green River’s course.

That link-between mantle behaviour and a specific river pattern-offers an unusually clear example of how “hidden” processes at depth can steer surface water and create distinctive landforms.

Rivers as fingerprints of deep Earth

The researchers plan to apply similar approaches to other rivers that cut across major ranges in North America. The goal is to test whether other apparently strange river alignments owe their origins to ancient drips or related deep processes, rather than to surface uplift alone.

Potential examples include rivers that slice abruptly across parts of the Rockies, or follow unexpected routes across the Colorado Plateau. If those river paths align with buried mantle anomalies, it could point to a broader pattern in how continental interiors change over time.

Process	Effect on surface	Impact on rivers
Mountain uplift	Raises terrain and steepens slopes	Encourages rivers to incise or divert around high ground
Lithospheric drip	Localised sag followed by rebound of the crust	Creates temporary low corridors that can redirect river courses
Isostatic rebound	Gradual uplift after weight is removed	Preserves established channels, often helping canyons deepen

Why the Green River story matters beyond Utah

Pinning down why a river runs where it does is not merely an academic exercise. River courses dictate how water, sediment and nutrients are transferred through landscapes, and they influence where people site dams, towns and reservoirs.

If mantle-driven changes can nudge rivers onto new routes over a few million years, the knock-on effects include altered flood tendencies, shifting erosion hazards, and questions over the long-term resilience of major infrastructure. Policymakers often plan in decades, but engineers responsible for very long-lived assets-large dams, or facilities such as nuclear waste repositories-sometimes have to think on far longer timescales.

Deep-earth dynamics act on timescales far beyond human planning horizons, yet they quietly set the stage for where rivers and canyons form.

There is also a climate angle. River incision and canyon development expose fresh rock, can influence local climatic conditions, and affect how carbon is locked up or released through weathering. Tying those surface patterns to mantle processes could improve models describing how continents and climate evolve together.

A further, often overlooked consequence is ecological: when a river entrenches and changes its sediment supply, it can reshape habitats downstream. Over long periods, such shifts may alter floodplain formation and nutrient delivery in ways that affect riverine ecosystems as much as they affect topography.

Key terms that help explain the “uphill” river

For non-specialists, several concepts underpin this story. They can sound technical, but they are central to why the Green River behaves as it does:

• Lithosphere: Earth’s rigid outer layer, made up of the crust and uppermost mantle, divided into tectonic plates
  
• Mantle: The hotter, more ductile layer beneath the lithosphere, where rock can slowly flow over long periods
  
• Isostasy: The principle that the crust sits in buoyant balance on the mantle, rising or sinking depending on loading and unloading
  
• Seismic tomography: A technique that uses the travel speeds of earthquake waves to map structures deep within the planet
  

With those ideas in place, the Green River’s supposed “gravity-defying” route looks less like a contradiction and more like a visible record of very slow, very large-scale rock movement beneath North America.

Imagining future landscapes shaped from below

Geologists routinely use computer simulations to explore how continental interiors might change over tens of millions of years. In these models, dense patches of lithosphere sag and drip while lighter regions rise, and rivers respond-like blue threads rearranging across a fabric that is gradually wrinkling.

If comparable drips occur beneath other parts of the western United States, valleys that currently carry modest streams could, over immense timescales, capture larger rivers. Canyons might widen, deepen or shift position. Places that seem stable within a human lifetime could look entirely different to a distant future observer.

For today’s outdoor enthusiasts floating through Lodore Canyon on the Green River, the landscape can feel timeless: sheer rock walls, cold water, and the steady pull of the current. This research offers another way to read the scenery-each bend and cliff as part of an ongoing negotiation between gravity, flowing water, and deep, unseen currents of rock far below the raft.