The more clearly the immensely speculative nature of geological science is recognized, the easier it becomes to remodel our concepts of any inferred terrestrial conditions and processes in order to make outrages upon them not outrageous.
—William Morris Davis,
“The Value of Outrageous Geological Hypotheses”
In 1984, shortly after landing a job at JPL, a geologist named Tim Parker found that he had a couple of weeks without too much to do. He decided to spend them in the image facility, looking at Viking photographs of Chryse Planitia. As part of a generation of geologists that had entered planetary science after Mariner and Viking, Parker was quite happy with the idea that vast floods had streamed into Chryse through the outflow channels to the south. What intrigued him was what had happened to the floodwaters after they left the channels. If they pooled as big, shallow, ice-covered lakes before draining into the permeable rocks below, those lakes would have had shores. If so, traces of those shores might still be visible. And if such shorelines were still visible, then Parker’s eyes might be peculiarly attuned to their discovery.
Today, North America is considerably better endowed with lake shores than it is with lakes. In the comparatively recent past—during the decline of Earth’s most recent Ice Age, between sixteen thousand to ten thousand years ago—the continent was home to many massive lakes that today have more or less completely vanished. Lake Agassiz, stretched out along the edge of the retreating Laurentide ice sheet that sat over eastern Canada, was at its greatest extent four times the area of Lake Superior today. Lake McConnell, to the north of Lake Agassiz, was more than six hundred miles long. On the south edge of the smaller Cordilleran ice sheet to the west was Lake Missoula, roughly the size of today’s Lake Ontario. And far from the ice sheets themselves, but fed by the rains that were part of the glacial climate system, were the great lakes of the West: Lake Manlius, part of the bed of which is now Death Valley; Lake Lahontan, in western Nevada; and, largest of all, Lake Bonneville, the shriveled remains of which persist today as Utah’s Great Salt Lake. Lake Bonneville has a particular place in American geological history because in the late nineteenth century it became the subject of the great monograph in which G. K. Gilbert deduced the lake’s past immensity by measuring its former shorelines—shorelines in the plural, because the lake disappeared in stages, leaving behind the distinctive rugby-shirt striping of a bathtub in a house full of students—and showed how the continental crust had risen up as the vast mass of water was removed.
In the early 1980s the shores of Lake Bonneville again became a site of geological interest. The Reagan administration planned to deploy a new generation of land-based nuclear missiles, the MX or “Peacemaker.” Unlike earlier American ICBMs, the Minutemen and Titans (whose descendants sent the Vikings to Mars), the MX was being brought forth into a world where the Soviet Union had demonstrated the ability to add new craters to the Earth with great precision and thus destroy missile silos in a first strike. Alternatives to old-fashioned silos were thus needed and one suggestion was that huge missile carriers—some loaded, some empty—could endlessly trundle between various sites in the western deserts on a specially developed network of roads or railways, thus confounding Soviet spy satellites. A great deal of surveying work was carried out in aid of this intercontinental shell game and Tim Parker was one of the surveyors, enjoying what for a geologist was almost heaven—a great deal of time in the field and terrific backup in the form of aerial photography. He spent two years tramping the shores of Lake Bonneville, learning to correlate the subtleties of lake deposits on the ground with the traces seen from the air.
Cloistered in JPL’s image facility, Parker started to look for shorelines in Chryse Planitia. Like those who had looked before, he found nothing. If anything, there were hints that the floodwaters had passed through Chryse without stopping and ended up farther north. North of Chryse were the even lower lands of Acidalia Planitia, where strange polygonal fractures in the ground that might reflect the drying up of watery sediments had been seen. But the Viking images of Acidalia were not so good, so Parker scuttled across the plains to the east. Here he studied a stretch of tablelands on the edge of Arabia Terra called Cydonia Mensae (the area where the “Face on Mars” had been seen). And there he saw traces of what looked like wave erosion around some elevated islands—islands like those whose shores he had walked and driven over in the Utah desert. Nearby he saw what looked like sand bars, another fossilized shore feature. Soon he was looking at a whole array of what he took to be shoreline features.
Having found possible shores, he set out to discover how big an area they enclosed. But the shorelines did not seem to want to meet up and define lakes. Instead, they obstinately headed north and east along the edge of Arabia’s cratered surface into the fretted terrain of Deuteronilus Mensae. There the new features changed their aspect, becoming sharper lines near the feet of the escarpments that mark the edge of the highlands, often facing each other across fretted channels. But they still looked like shorelines; just shorelines cut into a different sort of landscape. More worrying than the change in the features’ appearance was the fact that, east of Deuteronilus, they vanished. Parker had come to the blanket of debris surrounding a quite large and very spectacular crater called Lyot that sits just to the north of the northernmost part of Arabia’s great curve. On maps it looks like a marble that can’t decide whether to roll to the east or the west, or like one of the boulders perched on the rounded peaks of the Marabar Hills in A Passage to India. For hundreds of miles, there was no shoreline to be seen. And then, east of Lyot, in the area called Protonilus Mensae, the traces reappeared. At this point Parker started to get really excited. These shorelines were not going to close in on themselves any time soon. This was not the edge of a big lake. It was the edge of an ocean, one that must have filled most of the great northern basin.
In the following months Parker slowly traced his shorelines farther and farther around the highland-lowland dichotomy. The nature of the markings changed from place to place, as it had between Cydonia and Deuteronilus, but Parker knew that shorelines look different in different places. In some places the Viking pictures were not clear enough to pick up anything at all. In others, especially around Tharsis, the shorelines became obscured, quite possibly by overlying lava flows, and the thread girdling the ocean was lost. But it could always be picked up again later on. Eventually Parker traced the shorelines down the eastern edge of Tempe Terra, across the southern bight of Acidalia from Deuteronilus, until they finally faded out among the flood features of Chryse. What had started as a couple of weeks indulging curiosity had become the most laborious of circumnavigations, a trip Stan Robinson’s settlers would have been proud of. (In fact, Frank Chambers makes a similar trip in Red Mars, but in the opposite direction and only halfway around the planet.)
Like earthly oceans, the ocean Parker thought he saw often had more than one shoreline at any given place. This is not that surprising. Changes in the extent of continental ice cover mean that the Earth’s seas are surrounded by fossil shores above and below the current ones. During the next Ice Age, today’s shores will be anything up to a hundred yards above sea level. In some places on Mars, Parker saw half a dozen shorelines parallel to each other. Most of them, though, could not be traced for long distances. At the planetary scale there were just two: “Contact 1,” which kept close to the highland-lowland boundary, and “Contact 2,” which was farther out into the lowlands. Parker saw the two contacts—to call them shorelines in print would have been a breach of geological mappers’ objectivity—as evidence that the Martian ocean, like Lake Bonneville, had died in stages.
What Parker was seeing was dramatic enough to be slightly scary. At a time when the idea that Mars had significant amounts of water still required championing from people like Carr, to claim that there had been an ocean’s worth sitting on the surface would be highly controversial. Parker was by no means a senior figure—he had only recently joined JPL and as yet had no advanced degree. And his evidence was largely the evidence of his own eyes. Other people could look at the same Viking frames and see things very differently. Parker, a big but not overly self-confident man, whose loudness is revealed more clearly in his choice of T-shirts than in his blowing of his own trumpet, was not the sort to try to mount a revolution under those circumstances. He proceeded with caution. First he won over the man who had brought him to JPL, Dave Pieri, who had made the first catalogue of Mars’s valley networks in the 1970s. He then started work on his superior, Steve Saunders, a Mariner 9 veteran by this stage more interested in the Magellan mission he was planning to Venus than in Mars, and initially quite skeptical.
In 1986, Parker made a presentation on the theory toward the end of a symposium in Washington, D.C.; he overran his time—Parker’s details often overflow the limits of his presentation—and thus faced little questioning. Unfortunately he was followed by a man named John Brandenburg, also talking about an ocean but in a much more speculative way, and laboring under the disadvantage that he was also known to be an advocate of the Face on Mars, which the ocean would have made beachfront property (in Face mythology, which draws on the tropes of water shortage and lost civilizations in a reasonably predictable way, the impact that created the crater Lyot is linked to the loss of the planet’s atmosphere and ocean). Listening to Brandenburg, Parker recalls, “I was just sinking into my seat.” This was the sort of association he least wanted. For the next couple of years he kept quiet and worked on getting his Master’s degree.* When his interpretation of the features he saw as shorelines was finally published, in 1989 (by which time his boss, Saunders, was convinced enough to be a co-author on the paper), it was under the wonderfully anodyne title “Transitional Morphology in West Deuteronilus Mensae, Mars: Implications for Modification of the Lowland-Upland Boundary.” But the map on the second page showed the shorelines—sorry, “contacts”—stretching all around the great triangle of the northern plains. It was the sort of map people would notice.
One of those who noticed was Victor Baker of the University of Arizona. When Baker became a professor at the University of Texas in 1971 one of the graduate students there was Peter Schultz, the man who would later make the case for polar wandering to Mars. Schultz was getting access to Mariner 9’s pictures, and Baker, who had no background in astrogeology, found the newly discovered surface fascinating. He was particularly interested in the outflow-channel flood features—since quite by chance he was an expert on their closest terrestrial analogue, the “channeled scablands” near the city of Spokane.
The scablands lie on a gently sloping plateau in eastern Washington State, covering an area of about 12,000 square miles between the Snake River to the south and the Columbia River to the north and west. Across this plateau are great dark scars that widen and narrow to some hidden rhythm of the rock, joining and dividing like tangled hair. In places they are interrupted by streamlined islands; elsewhere they are crossed by long cliffs; their beds are grooved and in some places shaped into great long ripples, hard to see except from an aircraft, remarkably gentle in aspect until you remember that they are carved directly into the bedrock.
In the summer of 1922 the geologist J Harlan Bretz asked himself what could have made such a mess of this piece of basalt the size of Belgium. The answer he came up with was water, in very large amounts. By 1923, he was convinced that truly catastrophic floods had swept across the plateau, creating massive analogues to the features often seen in smaller streams—anastomosing channels miles wide, scouring by boulders instead of gravel and so on. Unfortunately, the uniformitarian ethos of geology made such an unheardof event a very suspect explanation: No one had ever seen such a flood. At a discussion in the Cosmos Club in Washington, D.C., in 1927, Bretz’s interpretation was denounced by many of his eminent superiors; few of them encumbered by any practical experience of the area. They all agreed that Bretz’s evidence was fascinating; then they all more or less asserted that the Brobdingnagian features must nevertheless have been made by repeated fairly commonplace floods of the Columbia River.
Bretz persisted. He sought ever more evidence but also took a stand of principle, invoking the ideas of William Morris Davis, one of the leading American geologists of the day, who the year before had had a paper published in the journal Science championing the role of the “outrageous hypothesis” as a spur with which to prod geology’s increasingly settled opinions. Bretz pleaded eloquently to his profession for a fair hearing:
Ideas without precedent are generally looked on with disfavor and men are shocked if their conceptions of an orderly world are challenged. A hypothesis earnestly defended begets emotional reaction that may cloud the protagonist’s view, but if such hypotheses outrage prevailing modes of thought the view of antagonists may also become fogged.
On the other hand, geology is plagued with extravagant ideas that spring from faulty observation and misinterpretation. They are worse than “outrageous hypotheses,” for they lead nowhere. The writer’s Spokane Flood hypothesis may belong to the latter class, but it cannot be placed there unless errors of observation and direct inference are demonstrated.
Over the next decade the fogged antagonists did begin to look for errors of observation and inference in Bretz’s work, even as Bretz himself tightened up the case. Then, in 1940, Joseph Thomas Pardee presented compelling evidence that toward the end of the most recent Ice Age Lake Missoula in eastern Montana had been drained at a truly spectacular rate when a glacial tongue of the retreating Cordilleran ice sheet had suddenly given way in northern Idaho. Bretz had never previously had a satisfactory explanation for the source of the floods and insisted that their existence should be accepted or rejected purely on the basis of the facts in the field—an in-your-face version of “what has happened, can happen” that, while logically coherent, did little to help his case. Now Pardee was giving him the perfect source: thousands of cubic kilometers of water suddenly unleashed through a valley just above the north end of the scablands. In his late sixties, Bretz went back into the field and, with two colleagues, wrote the definitive monograph on the subject, showing that there had in fact been a series of floods, after each of which the ice dam at Pend Oreille grew back, blocking the drainage channel and allowing Lake Missoula to refill itself. In 1979 the Geological Society of America gave Bretz its highest honor, the Penrose Medal; he was ninety-seven.
The citation that came with Bretz’s medal was written by Victor Baker. Baker had first come across the scablands when, as a schoolboy, he had made a plaster of Paris model of Washington State. A decade or so later, as a geology graduate student at the University of Colorado, Boulder, he took the Missoula floods as the topic for his dissertation, his first step to becoming Bretz’s successor as the acknowledged expert on the topic. When Schultz showed him the Mariner 9 pictures of outflow channels he immediately recognized some of the similarities and got in touch with Danny Milton, the member of Hal Masursky’s USGS team concentrating on the out-flow channels. Together, Milton and Baker wrote a definitive paper on the channels as flood features, a paper that beautifully drew out the similarities with the scablands feature by feature. It was what Baker calls an argument from coherence. While the various individual types of feature in the Martian channels were arguably open to other sorts of explanation, the flood hypothesis explained them all at once; it wove them together into a story.
His work on the Missoula floods shaped Baker’s future career as surely as water shapes a landscape. It introduced him to a strange form of scenery that he would go on to find in other parts of the Earth as well as on Mars. It taught him how to look at such landscapes and understand them, how to unify far-flung details into a single explanation through a tutored empathy for the force of the waters. In what was still a largely uniformitarian age the floods gave Baker a feeling for—perhaps a taste for—geological catastrophe. And his exposure to Bretz gave him his very own scientific hero to emulate, along with a profound respect for the power of the “outrageous hypothesis.” In 1990, he and a set of younger colleagues put forward one of their own, one that pulled all sorts of odd aspects of Mars—including Parker’s shorelines—into a coherent but previously unthinkable whole.
While everyone involved in this outrage would claim to have been led by what the data were telling them, in much the same way that Parker found himself whipped off on a trip all around the planet simply by following his eyes, it’s hard to talk to Baker for long without realizing that he had come to a stage in his life where an idea’s size and scope, not to mention its outrageousness, could be recommendations in and of themselves. Baker is a powerful speaker and a strong debater; he’s passionate about the value of geology and a vehement critic of any marginalization of his field by geophysicists and their like. He has devoted a lot of time to thinking about the differences in the way geology and physics see their worlds, the former through experience and imagination, the latter through abstraction and experiment.
“What makes geology strong,” Baker argues, “is the reality of its connection to the world, not the logic and structure of its thinking. In physics the connection to the world is tenuous because it only develops after you have produced your model and you’re trying to test it. Geologists immerse themselves first. The closest thing I’ve seen to this is the aboriginal people in Australia: They go out on the land and they have a sense that the land is their dreamtime story, a story in which these rocks and things are their spirit ancestors so everything is sacred but they know it intimately. A good geologist has the same sense of interrelationship and familiarity, a similar closeness of connection.” That sense of connection was what allowed Bretz to see the floods at Spokane and Baker to see the floods on Mars: understanding the landscape as a whole, correlating the features, making them cohere. Baker wanted to use the same sense at a planetary level, to develop a hypothesis that would explain a great number of disparate things not by appeal to geophysical modeling but through the sensibility of the geologist.
Jeff Kargel, a planetary geologist now with the USGS in Flagstaff, remembers watching Baker grope toward such a synthesis while Kargel was a grad student in Tucson. “I remember one of the most bizarre talks he ever gave was at a Mars conference in 1988 or 1989, a talk about a hydrological cycle that was mainly subsurface, a very strange talk, a very interesting talk, an unsettling talk. He was seeing that Mars is on one hand somewhat earthlike, but on the other hand it’s peculiarly Martian, unique, different from Earth, and this was puzzling Vic, just as it’s puzzled many other people before and since.” Not that long afterward, Kargel’s work helped Baker come up with a solution.
Although his doctoral work was on the outer solar system, Kargel was interested enough in Mars to have committed himself to looking at every single frame sent back to Earth by the Viking orbiters, using the university’s copy of the NASA photo archive. In early December 1989, he noticed a sinuous braided pattern in the southern part of Argyre, the basin in the southern highlands second in size only to Hellas itself. At first he thought it was a channel; then he realized that rather than being cut into the plain, the winding feature was raised above it.
Kargel’s undergraduate work had been in Ohio, a state once covered by the great Laurentide ice sheet; the Ohio River flows close to what was once its southern edge. From field trips there he was quite familiar with eskers, subglacial features formed when water cuts channels in the base of an ice sheet or glacier and deposits sediment on the rock below, leaving something a bit like an inverted plaster cast of an everyday river channel. Pressure exerted by the glacier, not gravity, drives the flow of the water, so eskers can run up slopes as well as down them. What Kargel was seeing in Argyre looked like a system of such eskers. Searching the neighboring frames, he found other features that might be glacial—sharp ridges that could be carved by ice, lumpy curved features that could be moraines. An argument from coherence was starting to form. Very excited, he talked about what he’d found to Virginia Gulick, another grad student, and one who was actually meant to be studying Mars. She was not convinced, but told him to talk to Baker, her supervisor. Baker quickly got excited; he pulled Kargel, Gulick, and a few more kindred spirits into a little impromptu seminar in the university’s Space Imaging Center. At the second of their meetings, someone brought along Parker’s innocuously titled shoreline paper and forced copies on everyone. Within days the outrageous hypothesis was taking shape.
You can’t build a glacier with water seeping up from underground; you need snowfall and so you need water transported through the atmosphere. The problem that Gulick was working on for her doctorate—a set of channels on the northern volcano Alba Patera that seemed to be caused by water erosion—also seemed to call for precipitation. Precipitation seemed to imply evaporation from some sort of sea or ocean. But the Martian ocean invoked by the warm-wet-early-Mars theory couldn’t have lasted long enough to have provided precipitation relatively late in Martian history, which was when the glaciers in Argyre and the channels on Alba seemed to have formed. For one thing, all the carbon dioxide would have been turned into carbonates long before. For another, the persistence of water in the atmosphere over such a length of time would have eroded away all the sharp crater walls in the southern highlands. So Baker suggested that the ocean was episodic. It came quickly, lasted long enough to explain what the grad students were seeing, then went away. And it did so repeatedly (hence the many different shorelines seen by Parker). For intimate blending of the earthlike and the alien, this idea has to score pretty highly; oceans are definitely earthlike, but transience is profoundly alien to their nature.
Baker’s story was based on the idea that the flow of heat from the Martian subsurface was highly irregular. On the Earth, convection currents in the mantle pump heat to the surface fairly efficiently. A lack of plate tectonics—at the moment, anyway—on Mars suggests that there’s no mantle convection and the heat has to move through solid rock by conduction, which is not a very efficient process (if you want to demonstrate this to yourself, get a long thin piece of rock and hold one end over a flame; you’ll put down the other end through boredom long before you put it down through pain). Baker suggested that such inefficiency would lead to an accumulation of heat in the mantle and that after a while—hundreds of millions of years, perhaps—the internal temperature would become high enough to force a huge reservoir of magma to the surface in a great volcanic burp.
This volcanic belching would contain a fair amount of gas, most of it carbon dioxide. This would thicken the atmosphere and produce a greenhouse effect. The vast amount of volcanism would heat up a large part of the crust, melting permafrost and opening aquifers, so water would spew out of all the outflow channels. Evaporation from these floods would add water vapor—also a powerful greenhouse gas—to the newly thickened atmosphere; the warmth of the waters’ passing would release carbon dioxide frozen into the plains over which they flowed. As the floods filled the great basin of the north, the surface temperature would climb above freezing. Within as little as a few weeks the waters would rise far enough to lap against Parker’s shorelines, their waves topped with fizzing foam as yet more carbon dioxide bubbled out into the warming sky.
From the moment the ocean was created, it would be living on borrowed time. Mars would simply not be able to produce a water cycle that could keep the ocean full or a climate that could keep it liquid. Carbon dioxide would freeze directly into the soils around the south polar cap. Carbonated water that fell in the south as rain or snow would seep deep into the porous rocks rather than running straight back to the sea along accommodating rivers, or being pumped back into the sky by transpiring plants, as happens on Earth. Carbonates would form underground and ice in the rocks would freeze. As the atmosphere grew thinner and drier, the temperature would drop until what remained of the ocean—much depleted by evaporation and by seepage into the porous rocks of its own bed—froze solid, forming the ground ice of the northern plains with its peculiar patterns and polygons. But the slow flow of heat out of the Martian interior would continue and, when enough heat built up for another great burst of volcanic activity, hundreds of millions of years later, the carbonated ocean frozen into the rocks would be recycled to the surface.
In this new view, Mars might never have been really warm, but it would have been intermittently wet through much of its history; the limits imposed by the lack of erosion in the south were circum-vented by cutting up a few million years of maritime climate and distributing the pieces over billions of years of history. Martian history, rather than being a matter of steady, slow decline, would have been dominated by slow underground cycles, the ocean’s brief appearances marking their visible crests. The cycles had begun in the Hesperian, it appeared, and continued until, well, in principle, today. Perhaps. But as the planet’s heart cooled, the wavelengths would grow longer; and with each cycle, something would be lost. Some of the carbon dioxide would be lost for good in the form of carbonates in the crust; some of the water would be lost from the upper atmosphere. And so each new avatar of what Baker and his colleagues came to call Oceanus Borealis would be smaller than the last, each new burst of maritime climate shorter and cooler.
In December 1989 its creators thought the episodic ocean did just what an outrageous hypothesis should. It explained a number of apparently quite separate anomalies, such as the valleys of Alba and the eskers of Hellas, the multiple shorelines of Deuteronilus and the details of morphology that suggested that the channels around Chryse, like their analogues in the scablands, had been used more than once. In a flurry of activity the Tucson team produced a set of abstracts for the March 1990 Lunar and Planetary Science Conference in Houston; they ended up presenting four papers back to back. “It upset a lot of people,” recalls Kargel. “It intrigued a few.” Parker was intrigued but skeptical—among other things, he saw the shorelines of a basin-bound sea in Argyre where Kargel saw moraines. Mike Carr was intrigued too—but perturbed by the way the theory threw around vast amounts of water with unaccountable abandon. The valleys of Alba might represent very easily eroded volcanic ash being eaten away by hydrothermal sources within the volcano itself. The evidence for glaciers was simply not convincing. And you just couldn’t get water out of the permafrost and aquifers quickly enough through the application of volcanic heat alone, let alone get the water back into the highlands with rain—enough rain to half empty the ocean—in a few millennia. Carr had a lot of disagreements with the people he began to call the Tucson Mafia; but he kept talking to them.
The Mafia had answers to some of the critiques, but on others they simply refused to be drawn. What they were offering, they said, was a story to be elaborated on and edited, not a model to be falsified. They didn’t know quite how the great episodic heatings worked, though they were happy when, a few years later, data from the Magellan mission made it possible to argue that on Venus volcanism came in great planet-wide spasms even more vast than the ones they invoked for Mars. (One of the Mafiosi, Bob Strom, played a key role in developing this interpretation of the Magellan data.) Nor could they say exactly where the extra carbon dioxide they needed was stored. But as far as they were concerned, their story of the episodic ocean’s source needed not to be precise, just to be plausible.
The heart of their case lay not in the details of heat flow, but in their geological feeling for the landscapes; and those landscapes, they argued, could only be explained by brief spates of watery climate relatively late in Martian history. This pattern of argument was familiar to all concerned from the controversy over outrages such as continental drift and the Spokane floods, where field geologists had pointed to things they could explain only by continental drift while others—not just physicists and geophysicists, but also other geologists—argued that the proposed cause was impossible. Over time, more and more disparate features were brought coherently into the picture; the objections were disproved or just dropped and the original geological insight was vindicated.
But for this to happen, new data were needed. Baker and Carr could argue until they were exhausted—there was a big set-piece confrontation at the American Astronomical Association’s Division of Planetary Sciences meeting in 1993, chaired by Carl Sagan—but there was no way for either to compel the other to his point of view. Geological history is full of such controversies, situations where people just see things differently. It is a byproduct of the fact that geology is ruled by perception and analogy, not experiment. If anything, planetary geology takes this aspect of its parent discipline’s nature to extremes—because the time-honored solution of going into the field and looking again, looking together, trying to see more, is just not available. All you can do is wait for the next mission.
* Parker’s caution was echoed by Baerbel Lucchitta in Flagstaff, who also started to think about widespread standing water in the northern lowlands in the mid-1980s, working on the basis of landforms within the lowlands rather than shorelines. Like Parker, most of the time she avoided the word “ocean.”