The Dust Detectives

Douglas Fox’s story, which covered the power of dust and its ability to shape ecosystems, won the Walter Sullivan Award for Excellence in Science Journalism in 2015. Fox is a freelance journalist.

Grass on the sand dunes dawdles in a breeze. The air drifting in from the Pacific Ocean is clear and cool on this gray February morning. But Kimberly Prather is not outside inhaling its salty tang. Instead, she stands in a windowless trailer parked behind the dunes, experiencing the air second-hand as it filters through a tangle of humming machines, tubes and wires. 

AGU’S Walter Sullivan Award for Excellence in Science Journalism

This award recognizes excellence in feature stories. This story was honored with an award in 2015.

“It’s just clean, clean sea spray,” she says, peering at a graph on a monitor. The lne is mostly flat, punctuated by several sharp peaks. “That’s sodium chloride,” she says pointing at one. Then, on another graph, she notes a set of jumbled peaks for silicon oxide, aluminum and sodium chloride. “This is dust mixed with sea salt,” she says — “a hint that the dust is coming from pretty far away.”

The air may seem clear at Bodega Head, 40 miles up the coast from San Francisco — but thousands of particles smaller than red blood cells drift in each cubic foot. Some have traveled a single mile; others, thousands. Prather’s machine sniffs them one by one, with a precision that no human nose can match. Its laser vaporizes five or six per second, spattering their guts across an ion detector, producing the chemical barcodes she peruses. A digital tally twirls alongside, like an odometer, too quickly to read — over 4,600 particles vaporized so far this morning.

Prather, an atmospheric chemist with the Scripps Institution of Oceanography at the University of California, San Diego in La Jolla, displays a similar frenetic energy as she narrates the second-by-second stream of chemical barcodes popping up on her machine.  “That’s a really good one, a lot of magnesium,” she says to me — then, into her phone, “Sorry, I’m showing him data.”

Cloistered in this compartment, glasses pushed back over curly brown hair, she is like a sibyl who can see far away, reading at a glance the stories told by each obliterated particle. She knows when a ship, invisible over the horizon, is burning dirty fuel by the trace of vanadium it leaves. Peaks for two isotopes of lithium, meanwhile, are “almost a ringer for Asia,” she says, indicating crushed coal incinerated by Chinese power plants. She knows these things because she has analyzed samples from hundreds of sources around the world.

Prather has toured the Western U.S. with her machines too, visiting coastlines, mountains, farms and even the suburbs of Southern California, where she chased semi trucks down Interstate 215. Her work on airborne dust and particulates could help answer a question critical to the drought-prone West’s future: What causes a cloud to drop rain or snow? And why does one cloud weep uncontrollably, while 20 others shed hardly a tear?

A cloud bulging above a mountain crest can hold several million pounds of water, but that water is divided into droplets too small to overcome the atmospheric updraft lifting them. Only rarely can liquid droplets grow large enough to fall on their own. Most of the time, they have to freeze into ice to achieve the kind of runaway growth — snowball, you might say — into something large enough to fall back to Earth as rain or snow. But freezing is harder than you might think. Never mind the 32 degree Fahrenheit freezing point that you were taught in school: Tiny water droplets can remain stubbornly liquid at temperatures as low as 20 or 30 degrees below zero, perched on a thermodynamic cliff.

Scientists have long believed that the instigator that tips them over the edge, into becoming ice, hides somewhere within the microscopic flotsam floating in the air. They have searched for it for decades. Now, Prather and her colleagues are finally closing in on its identity.

Ironically, this new information about the tiniest particles is changing our view of how the world works on a grand scale, revealing a vast network of airborne dust commerce that ties together far- flung parts of the globe. These connections have enriched the West’s soils and shaped its precipitation patterns for thousands of years. But human-driven global warming and pollution have already begun to alter them, with implications for everything from agriculture and drinking water supplies to the basic shape of Western ecosystems. “We don’t understand this complex, totally intertwined system,” says Prather. “But we need to understand it soon.”

If you’ve read the novels of Kurt Vonnegut, then you’re already familiar with the notion that ice nucleation has profound power to shape the environment; in Vonnegut’s 1963 Cat’sCradle, a mad scientist creates a substance
called ice-nine that seeds ice formation at
temperatures up to 114 degrees Fahrenheit,
leading to an apocalyptic moment in which ice spreads through all of the world’s oceans, rivers and lakes. Vonnegut was doing PR work for the General Electric Research Laboratory in Schenectady, New York, when he came across the idea: Irving Langmuir, a GE scientist, had mentioned it during a visit with H.G. Wells years earlier. But the novelist may have had another inspiration, as well — his older brother, Bernard Vonnegut, who worked as a cloud physicist at the same lab.

The elder Vonnegut began studying ice nucleation in 1946, in hopes of manipulating weather and preventing airplanes from icing up and crashing when they flew through supercooled clouds. In an artificial cloud chamber, Bernard Vonnegut and his colleagues tested the ice-seeding abilities of hundreds of substances, from powdered graphite to oil to tobacco smoke. Silver iodide, he discovered, was extremely potent. When he blew gently over a red-hot silver wire into a chamber filled with supercooled water droplets and iodine gas, his breath coalesced into a scintillating swirl of ice crystals, produced by silver atoms that wafted off the wire and reacted with the iodine.

Bernard Vonnegut saw silver iodide’s geometry as the key. The pattern of positively charged silver atoms and negatively charged iodine atoms, stacked into a cube-shaped crystal, perfectly matched the spacing of water molecules’ hydrogen and oxygen atoms in an ice crystal: Water molecules could stack onto the salt like Legos, starting ice formation.

Vonnegut’s discovery, published in 1947, helped start a scientific revolution. By the early 1950s, California and several other states had started cloud-seeding programs, using airplanes or ground-based machines to spew silver iodide into the sky.

Those programs produced uncertain results over the ensuing decades. And since silver iodide was manmade, the search for the natural ice seed continued. Candidates over the decades have included everything from tiny diatom shells to clay minerals to extraterrestrial dust. But even when laboratory experiments suggested that a particular substance might be good at making ice, scientists found little evidence that it did so in actual clouds. The question was still open when Prather began drifting toward it, unwittingly, in the early 1990s.

After several years working under Nobel Laureate Yuan T. Lee at the University of California-Berkeley, studying how large molecules disintegrate when blasted with lasers, Prather decided to set her sights on a problem more immediately relevant to society: Pollution. Doing this would require building scientific instruments that simply didn’t exist at the time — and she was fortunate to have help from an unlikely source. While earning her doctorate, she had become close friends with Joe Mayer, a machine-shop tech at the university. Mayer’s father once fashioned models of military aircraft wings from mahogany for NASA to test in its wind tunnels. Mayer and his brother later started a machine shop of their own in the family’s barn, where they built commercial fishing gear, tractor and plow parts, and the occasional NASA satellite component.

Over the next several years, Prather and Mayer built a machine designed for her new course of study. Instead of firing a laser into a cloud of vapor (essentially, shooting fish in a barrel) as she had been doing, it used vacuum nozzles to focus chaotically drifting particles into a single-file beam traveling near the speed of sound, allowing them to be picked off one by one as if by a sharpshooter. It didn’t work perfectly — “we hit about 30 percent of them,” says Prather — but it was better than anyone else had done.

In 1992, the pair moved to U.C. Riverside as husband and wife. “If you find a good machinist,” she says, “you should marry him.” By 2002, they’d built a smaller machine called Laverne that could fit in the back of a van; five years later, they built Shirley, which was petite enough to fly on airplanes.

Prather drove these machines through Santa Ana dust storms, across the smoggy San Bernardino Valley and past power plants — measuring pulses of pollution from moment to moment, watching as grains of soot, sea salt and dust became slicked-over with nitrates and sulfates produced by fossil fuel burning. Now and then Laverne was shipped off on research vessels in the Pacific Ocean to measure pollution drifting from Asia.

In 2007, Prather and Shirley were invited aboard an airplane that was sampling rainclouds over the sagebrush plains of Wyoming. Its menagerie of scientific instruments would analyze air sucked in from a tube on the wing. It was Prather’s first chance to sample the inside of a living, breathing cloud. The results of that flight would change not only the course of her career, but also that of ice nucleation research.

As the C-130 Hercules soared at altitudes up to 26,000 feet, Prather watched Shirley’s laser click on and off, and quickly saw a pattern in the graphs. Every time the plane entered an icy part of a cloud, prominent peaks for iron and titanium appeared. And the second it exited, those peaks disappeared. It was a scientific first — correlating the second-by-second presence or absence of ice with the composition of airborne particles.

Those metallic particles were clearly mineral dust, since they also contained silicon and oxygen, but they weren’t local. Their origin might have remained a mystery had their chemical makeup not reminded Prather of data she had gathered 13 years before from a ship off the coast of China. A review of satellite weather data confirmed her hunch: The strange dust appeared to have come from a windstorm eight days earlier over East Asia.

A month spent flying over California’s Sierra Nevada in 2011 corroborated this: Icy parts of a cloud almost always contained the same iron- and titanium-rich dust. The amounts were minuscule — roughly a quarter-cup of material sprinkled through a cubic mile of cloud. But this dust was potent enough to seem magic, able to seed ice at temperatures up to minus 5 degrees: nature’s version of the manmade salt that Bernard Vonnegut had discovered 60 years before. The dust that spawned ice and rain over Wyoming and California had drifted 8,000 miles, from a desert in China that few Americans have ever heard of.

Bactrian camels during a storm in the Taklamakan; dust from this desert can rise into the jet stream and be carried around the globe. Courtesy of Aaron and David Putnam.

Three of the world’s tallest mountain ranges bristle across the remote northwest corner of China — the Pamir, Tian Shan and Kunlun, whose peaks rise to 24,000 feet. Squeezed up from miles beneath the Earth’s surface by the collision of tectonic plates, their ragged granite and sandstone edges are slowly succumbing to cycles of freeze and thaw, and to microbes that oxidize traces of iron, reducing the rock into gravelly debris that sloughs off the mountains’ weathering epidermis. Over millions of years, glaciers have pulverized this mineral dander into a powder as smooth as cream cheese, and winds and intermittent rivers have spread it over the Tarim Basin. There, caught between the mountain ranges with no outlet to the coast, it has created an inland sea, not of water, but of sand and dust — an expanse of migrating dunes second in size only to the Sahara — The Taklamakan Desert, often called the Sea of Death.

Centuries ago, the great Silk Road forked to avoid this vast desert, swinging north and south to its edges. Even today, the Taklamakan extends tendrils of invisible influence into
faraway lands. Dust blown from other nearby
deserts, like the Gobi, stays closer to the ground,
often settling within a few hundred miles and
compelling residents of Beijing to shield their
faces with breathing masks. But dust from the
Taklamakan travels much farther, owing to an
accident of geography and atmospheric
circulation. Each spring, winds rake the desert’s
surface, forcing dust up to 25,000 feet as it
billows over the Tian Shan Mountains to the north. At that height, it enters the jet stream over Siberia, which carries it thousands of miles east. The Taklamakan and other Asian deserts send up to 400 million tons of their mineral wealth blowing over the Pacific each year. Much of it simply falls in the ocean: Sediment cores suggest that up to 60 percent of the mud on the seafloor comes from Asian deserts. Some falls inside raindrops or snowflakes over North America. Some completes an entire circuit of the globe before it falls.

Scientist Aaron Putnam amid the blowing dust in China’s Taklamakan Desert. Courtesy of David Putnam.

The process is part of a great invisible web of global dust migration that also includes the Sahara, the Arabian Peninsula, and deserts in India, Afghanistan and North America. But dust from the Taklamakan carries special significance for the Western U.S. Somewhere over the ocean, plumes of this traveling “aerosol,” as scientists call it, sometimes collide with an “atmospheric river” composed of water evaporated from the equatorial Pacific, hundreds of miles to the southeast.

Twenty or so atmospheric rivers drift over the West Coast each year. They generate a handful of major storms that together supply nearly half of the rain and snow that falls on California, 30 percent of the precipitation that falls on Oregon and Washington, and much of what falls on the inland West as well.

These atmospheric rivers vary widely in how much moisture they produce. Some drop 25 percent of their water; others, just 15 percent. “This is where the aerosol comes in,” says Marty Ralph, a meteorologist who led water-cycle research at the National Oceanographic and Atmospheric Administration in Boulder, Colorado, for 13 years before moving to Scripps a couple of years ago. He and Prather recently teamed up to compare two atmospheric rivers that passed over California. They had nearly identical water content, temperatures and winds. But only one contained Asian dust — and it dropped 40 percent more precipitation than the other. That corresponds to a staggering 1.5 million acre- feet of water — greater than the amount of water currently sitting in California’s largest reservoir, Lake Shasta.

That Asian dust triggers Western rain is a momentous discovery. Still, a major question remains: What exactly is it about this dust that triggers ice formation?

The simplest explanation is that specific kinds of mineral crystals provide a good scaffold for water molecules to latch onto, as with Bernard Vonnegut’s silver iodide. Experiments published in 2013 show that one class, called K-feldspars, trigger ice especially well. Although K-feldspars comprise only a minor component of dust worldwide, they are consistently found in dust from the Taklamakan Desert.

There’s likely more to it, though. Hitchhikers on the dust’s surface may play a key role. Back in the trailer at Bodega Head, a set of conspicuous peaks for carbon-nitrogen and carbon-nitrogen-oxygen catches Prather’s eye. “A biological marker,” she says, “a piece broken off a protein” — most likely the incinerated innards of a living cell that had been riding on the speck of dust she just zapped.

Forty percent of the Asian dust she’s sampled over Wyoming and California contains such signs of life. Other scientists have found that desert microbes can survive transoceanic transport on grains of dust. The same traits that allow them to inhabit hot deserts also provide protection against the aridity and ultraviolet radiation of the upper atmosphere, an advantage that may allow them to ride the global dust train from one continent to new habitats on another.

A few microbes even secrete proteins that nucleate ice — some even more potently than silver iodide. And scientists collecting ice-nucleating particles from snow in Montana have found that heating them to bacteria-killing temperatures or treating them with a bacteria-obliterating enzyme — processes that don’t alter mineral structures — sharply reduces their ability to create ice.

But other, tinier vermin could be even more important — hitchhikers on the hitchhikers. Sometime after noon at Bodega Head, Prather leans close to inspect a barcode with a towering peak of phosphorous — a classic signature for something containing lots of DNA and not much else. “Oh, my goodness,” she says. “That was the viruses.”

Any dust that is carrying bacteria is just as likely to carry these ubiquitous genetic parasites. And while bacteria have squishy membranes that may provide poor footing for water molecules — like stacking Legos on a waterbed — many viruses are sealed in rigid protein cases. Those cases are covered in a repeating tile pattern of positive and negative charges that water molecules might just latch onto with ease. “Maybe,” says Prather, “that pattern is magical.”

David Putnam documents ancient stumps in the Taklamakan Desert. Courtesy of Aaron Putnam.

Unveiling the true ice-maker could have important implications, especially for the American West, where the water supply has always been fickle. Knowing nature’s recipe for ice could help improve cloud-seeding programs, which still exist in most of the region’s states, despite their dubious effectiveness. It could help atmospheric scientists predict whether certain types of pollution might either disrupt or enhance the formation of ice.

And knowing the geographic dust reservoirs responsible for precipitation could also allow researchers to predict what might happen as climate change begins to alter wind patterns — and hence, dust migration. Due to a quirk in fluid mechanics, the ability of wind to transport dust varies with the cube of the wind’s velocity: Double the wind’s speed and you increase dust transport eightfold; cut the speed in half, and you reduce the transport of dust by 88 percent.

Dust transport and deposition turn out to be exquisitely sensitive to changes in moisture, as well. The Taklamakan Desert, for example, has not always been a prolific dust-maker. The origin of the desert’s name is highly disputed; some say the Sea of Death — but others give it more mysterious roots — derived from Turkic or Uighur phrases for a “place abandoned” or a “land of poplars.”

In 2010, David Putnam, an archaeologist from the University of Maine in Presque Isle, and his son, Aaron Putnam, a geologist from the Lamont-Doherty Earth Observatory in New York, came across something strange while exploring the desert’s sandy expanse: the stumps of poplar and tamarisk trees, some of them three feet across, rooted in the deep troughs between dunes. They found the shells of aquatic snails in these small patches of exposed ground, and layer upon layer of dry mud. Elsewhere in this same desert sits Niya, an ancient outpost of the Silk Road where the splintered remains of wooden houses and apricot orchards still protrude from the dunes — one of a dozen abandoned towns now buried by sand. The message is clear: The Taklamakan has been quite wet at certain times in its recent past, sharply reducing its dust output.

Even without knowing when exactly those wet times occurred, it leads one to consider the decades-long megadroughts that the West has suffered several times in the last 2,000 years. Could fluctuations in Asian dust storms — their onset or cessation — have influenced those droughts’ intensity? Changes in dust are often considered a symptom of environmental change rather than its cause, but scientists are beginning to believe that the shifting migrations of dust might actually propagate or amplify environmental changes. And those changes aren’t limited to water.

The Southwest’s Colorado Plateau is built of sandstone notoriously poor in mineral nutrients. It is the dust blowing in from the Mojave and other Western deserts that nourishes its vast juniper-piñon woodlands, blanketing the mesas and valleys in layers up to four feet thick. And 11,000 feet up in southwest Colorado’s San Juan Range, dust provides 50 to 80 percent of the soil’s mineral nutrients — not only sodium, potassium, calcium, and phosphorus, but also micro-nutrients such as copper and zinc. “The reason you have all of these incredible flower-covered alpine meadows is because you’re putting dust into these places,” says Jason Neff, the biogeochemist at the University of Colorado in Boulder whose team collected this data.

So dependent is the West on its lifeline of dust, its ecosystems have actually evolved to help capture it. Drive Highway 211 in eastern Utah, and the desert plain seems monotonous and flat. But a closer look reveals that the soil surface is corrugated into silver dollar-sized polygons speckled a mildewy black.

This living fabric represents an archaic throwback to the microbial mats that first colonized Earth’s barren land 2 billion years ago — long before the advent of plants. Its mishmash of cyanobacteria and lichens cannot grope around in the dirt for nutrients as modern plants do with roots. Instead, these biological crusts, as scientists call them, capture mineral nutrients from the air on sticky polysaccharide threads secreted by the cyanobacteria. These crusts may once have covered most of the West, gathering up to 100 billion tons of nutrient-rich dust since the last ice age — anchoring it against the prying fingers of the wind and building it into soil.

Today, that process is reversing. Levels of circulating dust have increased by fivefold in the last 150 years, as road-building, farming, energy development and the hooves of cattle crush the crusts and release their long-stored larder of minerals to the winds.

This process is spurring yet more changes.
Dark, heat-absorbing dust sprinkled on the
snowpack of the Rocky Mountains is
accelerating spring melt, a trend that might one
day reduce the amount of water available for
agriculture and ecosystems alike. Meanwhile,
nutrient-rich dust settling on Wyoming’s Wind
River Range is causing high lakes “in basins
with no vegetation, no trees, no soils,” to bloom
with algae, says Neff, tinting their once crystal-
clear waters green each summer. Neff’s doctoral
student Janice Brahney, now at the University of British Columbia, has even found that increased dust is also changing the chemistry of the rain, as alkaline minerals neutralize nitric and sulfuric acid produced from burning fossil fuels.

The fact that mineral dust reacts with nitrate and sulfate to neutralize acid in a cloud droplet may sound like good news, but there’s a caveat. Researchers have shown that contact with acid-forming nitrates and sulfates can sharply reduce the ability of some types of mineral dust to trigger ice formation — with potentially ominous effects on precipitation.

That’s because the chemical reaction that neutralizes the acid can also leach out the mineral’s potassium or sodium atoms. “You can think of it as etching a surface differently,” explains Paul DeMott, an atmospheric chemist at Colorado State University in Fort Collins who frequently collaborates with Prather. This disrupts water molecules’ ability to assemble into ice crystals. A mere five hours of contact with air pollution can alter dust in these ways, and evidence now suggests that K-feldspars, the most potent ice-forming minerals discovered to date, are also among the most sensitive to this acid etching.

Fortunately, Taklamakan dust has little contact with pollution — at least for now. Its journey over the Tian Shan Mountains carries it over only sparsely populated areas before it reaches the Pacific. The pristine state of this dust may well be part of what makes it such a potent rainmaker in the American West. But Prather worries that this could change as emissions from both Asia and the West Coast increase. Her instruments sometimes detect Asian pollution drifting to the West Coast, where it could conceivably mix — and react — with Asian dust. Prather has seen this during the month of February, a vulnerable time given that atmospheric rivers, and the rain they bring, tend to arrive in California between October and February. “We only have a few months to get all of our rainfall,” she says. “So if we’re impacting even just one of those months, that’s important.”

Water supplies in the West may be sensitive to the effects of pollution for another reason, as well. Most of the region’s water derives from a process called orographic precipitation, in which clouds drop rain or snow during a short time window as they rise and cool over mountain ranges. A cloud droplet swept over the Sierra Nevada may have only 30 to 60 minutes to grow, freeze and drop. After that, it sweeps down the far side of the range, warms and re-evaporates. Even minor changes in the rate of droplet growth or ice formation can throw off the timing of this delicate choreography, and, potentially, reduce precipitation. Daniel Rosenfeld believes that increased particulate pollution is already causing this to happen in subtle ways.

A cloud physicist at the Hebrew University of Jerusalem, Rosenfeld has spent 25 years studying precipitation in the Western U.S., whose arid lands resemble Israel’s. Since 2003, he has analyzed records from hundreds of rain and snow gauges, comparing precipitation over mountain ranges that were downwind of either unpopulated territory or large urban pollution sources. In the latter case, precipitation appears to have decreased by up to 30 percent over the rising hills of the mountains, and sometimes by a few percent over their crests, while precipitation on the downwind dry side of the range actually increased.

Rosenfeld blames this change on ultra-fine soot, generated from the burning of fuel and biomass, that rises into the clouds. These particles increase the number of water droplets in a cloud, but each droplet is far smaller. At one ten-thousandth the minimum size needed to fall as rain, they don’t collide as easily as larger ones, so they grow more slowly. But as they rise higher over the mountains, a backup process kicks in: The falling temperature finally spurs ice formation. Hence the polluted cloud may end up dropping just as much water — but the precipitation is delayed, shifting farther downwind.

Changes in the geographic distribution of precipitation could eventually necessitate the building of new storage reservoirs in downwind basins, or relocating farms to follow moving water supplies, says Ralph: “Where the water falls really matters.”

Several droughts have struck the West in recent years, most recently in California, where water storage in reservoirs now sits near 50 percent of its season-adjusted average. Up at Lake Shasta, rusty, silt-coated trestles have emerged from the receding water: abandoned railroad bridges that spent decades as much as 100 feet below its surface. The California drought stems in part from shifting storm tracks, which sent the state about half as many atmospheric rivers as usual last winter. Global warming is expected to worsen this problem by pushing the planet’s desert belts and storm tracks farther toward the poles. Changes in Asian dust could exacerbate this problem by causing each atmospheric river that still does reach the West to drop less water. It is against this urgent backdrop that Prather and her colleagues continue their search for the ice seed.

DeMott, one of those colleagues, studied as an undergraduate under Bernard Vonnegut in 1978. The kindly old professor rode a single-speed bike to work each day and taught him how to create an artificial cloud inside a 10-gallon water bottle. DeMott spent a decade developing a machine that picks out the rare particles able to trigger ice. Rather than blasting them and sniffing the smoke, as Prather does, he has collected them into a vast library of ice-making specks — roughly 50,000 to date. He plans to subject some of them to DNA sequencing, to identify any microbial passengers. Those microbes, or the viruses that feed on them, could then be tested for their ice nucleation properties.

Unlocking the particles’ secrets will require exploring tiny frontiers that exist only within a fleeting sliver of time and space. An ice crystal’s moment of conception is tenuous, with as few as 70 water molecules clinging to the surface of a solid object. Some are bucked off by thermal vibrations, while others grab hold to take their place. The nascent crystal grows slowly at first, but after a hundredth of a second, it reaches an invisible threshold where its size confers stability. From there it grows exponentially — like ice-nine, the apocalyptic tide of frost that overtakes the globe in Kurt Vonnegut’s novel — snapping the entire droplet into ice.

The catalyst of this chain reaction may measure less than one 10-millionth of an inch across: If a grain of Taklamakan dust were the size of planet Earth, the ice-nucleating site on its surface might be no larger than a medium city, like Tucson or El Paso. And so DeMott has turned his attention to the molecular detail of individual ice-seeding particles amassed in his vast archive. Using spectroscopy and other methods, he maps the chemical surface of each particle — a process not unlike mapping the surface of a newly discovered planet: water here, craters there, plains and mountains here and there. Just 150 years ago, European explorers wandered deep into Africa, looking for the source of the Nile. Today, that quest plays out on a grain of dust. Somewhere within that minuscule world hides the source of ice, rain and snow — the source of all rivers.

© Douglas Fox 2018.

 You may read this story in its original format and view additional multimedia on High Country News