Letters From Grand Canyon—Piracy and Capture Carve the Grand Canyon: Part B

Two issues ago, in bqr Volume 15 Number 3, we ran the first part of this article. Here continues the geological story of the area of northern Arizona and the age and formation of the Grand Canyon. This is a complicated story with many theories and countertheories.
Conflict Resolution II: Another New Concept
At this point, we seemed to be back to square one—an ancestral river more or less traceable to near the east side of the Kaibab Plateau, but no further. A way out of the dilemma was suggested to me by work in northern Arizona, especially on the Shivwits Plateau. The key was the presence of lava remnants as old as nine million years that protected details of the ancient landscape over which they flowed. This made it possible to describe what the landscape of northern Arizona looked like at various times in the past nine million years, and to work out how the erosional processes through the landscape evolved with time. Both are important parts of the Grand Canyon story.
The Lava Story: Ancient Valleys
The southernmost part of the Shivwits Plateau is a long finger that points south and is surrounded on three sides by the western Grand Canyon. The finger is capped by basalt lavas eight to six million years old. Today, you can stand at the eroded edge of these lavas and look dizzyingly down onto the precipices, fretted buttes and cascading drainages that are part of the western Grand Canyon, but when they were emplaced, the lavas flowed over smooth terrain with no trace of dissection; what’s more, there is no evidence that the lavas cascaded into the canyon, as do the famous younger ones near Toroweap, and the older ones on the Grand Wash Cliffs. The conclusion is clear: the western Grand Canyon did not exist eight to six million years ago when the lavas flowed over the land.
Lava remnants scattered over northern Arizona, including the Shivwits Plateau, tell us a great deal more. Most of this region is underlain by the Kaibab Limestone of Permian age, the rimrock of the Grand Canyon. The lavas, however, do not rest on this formation, except for the very youngest ones. Instead, they rest on the Moenkopi Formation, the next geologic unit above the Kaibab. The Moenkopi is soft and easily eroded, but lavas are hard. Therefore, areas where lavas once flowed over the land have been protected, showing us what the surface of the land was like at the time.
What we learn is this: first, northern Arizona was floored until quite recently by the red Moenkopi Formation, not the gray Kaibab Limestone; think of the country near Wupatki to get a mental picture of this landscape. What is more, the land surface was considerably higher than it is today. In any given area, the older the lava, the higher it is above the present surface, and the greater the thickness of Moenkopi—as much as about 1000 feet—beneath it. There is another curious fact: for lavas of the same age, those farther northeast are highest and have more Moenkopi beneath them. The shape and position of the lava remnants leads us to an explanation of this fact.
Molten lava flows much like water, seeking the lowest spot and eventually making its way down valleys, as river water would. Lavas that flow down valleys should have an elongated shape; this is just what many lava remnants in northwest Arizona show. The elongation is in a generally northwest-southeast direction, and many of the vents from which the lava flows issued are at the southeast end of the remnants. We can conclude that a common landscape feature of northern Arizona several million years ago were valleys that trended northwest and sloped in that direction. The floors of these valleys are now preserved under the lava remnants well above the present surface of the land. Lavas that flowed into the valleys from vents on the valley sides are the ones that show us what the valley sides looked like.
Scarps and cliffs Move Across the Land
Was there a particular set of geologic circumstances that controlled the location of these ancient valleys, perhaps a particularly soft layer into which valleys could easily be carved? Indeed there was. But to understand this, we must first think a little about how the landscape of northern Arizona has evolved. The controlling factors here are two: the composition of the sedimentary rocks, and the slope of the rocks.
Much of northern Arizona is underlain by Mesozoic rocks, the colorful strata typical of the Colorado Plateau, the ones that, in the geological layer cake, are above those forming the Grand Canyon. Not long ago such rocks formed the surface of nearly all the Plateau, including the Grand Canyon region. These strata consist mostly of soft sandstone and shale, interrupted in places by more resistant layers such as the Shinarump Conglomerate, the cream-colored pebbly layer so prominent near the boat-launching ramp at Lees Ferry. The resistant layers protect the softer ones underneath from erosion, forming a cliff or scarp. The maroon cliff north of the road to Lees Ferry is an excellent example of such a scarp.
Do these cliffs remain in place once formed, merely getting fretted and worn down with time? They do if the strata are flat-lying. The results of this process are well displayed by buttes and mesas of Monument Valley. But if the strata are not flat-lying—if they have a dip—a very different process takes place, a process that is typical in northern Arizona, where the beds have a very gentle dip to the northeast.
To begin with, it is an observational fact that the cliffs face updip, that is, up the slope of the beds. In northern Arizona, this means the cliffs face south or south-southwest; think of the Vermillion Cliffs or the Grand Staircase of Utah. But it also happens where the strata are disturbed by a more local feature such as a fold. The Kaibab Plateau is a big dome-like fold, and the cliffs face up the dip created by this fold. You can see this along House Rock Valley and at the circle cliffs cut by US 89 at the north end of the Kaibab.
The second component of the process is that the cliffs and scarps slowly and majestically retreat downdip with time, because the cliff faces are attacked by especially intense erosion whereas the mesa behind the face is not. If you could make a movie of the landscape that shows what happened over the last nine million years or so, you would see just that—cliffs and scarps retreating slowly northeast over wide areas, and off a fold like the Kaibab at the more local scale. Not even Hollywood can make such a movie, but our friends the lavas have done it for us—if you learn to look at them with a geologist’s patient eye.
It so happens that the favorite place for valleys to form during the eight million-year interval documented by the lavas is at the foot of the Moenkopi-Shinarump scarp. Earlier lavas, had there been any, probably would have documented valleys formed higher in the geologic section, for example in the weak Chinle Formation at the foot of the Vermillion Cliffs. The lavas that did flow over the land show that the ancient valley at the foot of the Moenkopi scarp was wide and gentle—and this is just what the modern valleys in this geologic position look like today. You can see one at the north end of the Kaibab as you drive toward Kanab. For the modern valleys in this region, we know that the Shinarump-capped scarp that forms the north side of the valley is at most a few miles from the valley floor. The same holds for the ancient valleys.
Remarkably, lavas of different ages always occupy a valley in this geologic position—the foot of the Shinarump scarp—but in different geographic locations, depending on the lava’s age. About eight million years ago, the valley was near Mt. Dellenbaugh; about four million years ago, near Poverty Mountain; 1.4 million years ago, a few miles northeast of Wolf Hole; about one million years ago, near Clayhole Wash. This tells us that the Moenkopi-Shinarump scarp has been retreating northeastward at the considerable rate of two to three miles per million years—faster than some people walk. So, hills are not eternal, landscape is anything but immutable, and we cannot use the landscape of today to make guesses about what might have been going on even a short time in the past if we don’t have information like that provided by the lavas. You grasp what this means quite vividly when you realize that the Vermillion Cliffs, today near Kanab, most likely were near the Grand Canyon (in the area of the present Whitmore Wash) some six million years ago. Furthermore, the rocks of the Vermillion Cliffs extended right over the Kaibab Plateau until just a few million years ago. How did this happen?
Crossing the Kaibab Plateau
Having created a mental movie showing the evolution of landscape in the Grand Canyon region, we can return to the ancient Colorado River, which we had left stranded on the east side of the Kaibab Plateau, with no obvious continuation. Now we can take a stab at figuring out what the landscape that the ancestral river flowed through might have looked like.
First, we know that the characteristic feature was valleys trending northwest and bordered by scarps on their northeast side. Many of the valleys drained northwest. Do we have valleys suggestive of this pattern now? Yes: those of the Little Colorado, Cataract Creek, and Kanab Creek. Near their confluence with the Colorado River in the Grand Canyon, these valleys are as narrow and rugged as other tributary canyons. This is the young part of the valleys, modified by rapid downcutting by the Colorado River. Away from the River, however, the valleys are long, well developed, and “mature”, contrasting greatly with all other washes tributary to the Grand Canyon, which are short and precipitous. The valleys are left-overs from the time before the Grand Canyon was formed.
Second, we must restore the land surface to its former higher position, maybe even thousands of feet higher than it is today. We do this by mentally putting back in place the Mesozoic formations, some softer, some harder, that were still present at the time of the ancestral river but have been eroded since.
Having done these mind experiments, we discover something quite remarkable: it is very likely that the top of the Kaibab Plateau, which corresponds to the axis of an up-arched fold and which seemed such a formidable barrier for the river, may have been topographically lower than the sides of the fold at various times in the past. Today the Kaibab Plateau is a huge whale whose back towers thousands of feet above the surrounding terrain, so you may wonder how anyone could suggest the absurd reversal of topography that I am proposing. To see that the proposal actually makes sense we have to play the erosion movie backwards. An excellent place to do so is at House Rock Valley north of US 89A. The valley has a northerly trend. Its east flank is a continuation of the Vermillion Cliffs; the west side is the Kaibab Plateau, whose strata slope down toward the valley. Let us now remember the movie that shows how cliffs retreat with time down the slope (dip) of the rock layers. If we play the movie backwards, we can bring the cliffs gradually up the slope of the Kaibab to the very top of the Plateau. Here, the cliffs moving upslope from the east meet those that have similarly come up from the west. These are the cliffs that now form part of the Grand Staircase near Kanab. Clearly, these rocks used to extend smooth and unbroken right over the Kaibab Plateau. There were no Vermillion Cliffs here at the time. However, hard-over-soft couplets of strata higher in the geologic section formed other cliffs east and west of the axis of the Kaibab Plateau. This particular frame of the geologic movie bears close scrutiny because it has much to say about the doings of the ancient Colorado River.
On the crest of a fold, rocks are fractured by the bending that formed the fold. The fractures make the rock more susceptible to erosion. The consequence is that any stratum that is at the topographic surface because erosion has removed overlying strata will be subjected to accelerated erosion along the crest of the fold, and will be cut through there first, forming a depression bounded by slopes or cliffs on each side. The crest of the fold is now lower topographically than the sides. Where a fold plunges (you can see this along the north side of the Kaibab Plateau), cliffs formed by resistant layers, and valleys formed by the soft ones describe curving patterns abundant on the Colorado Plateau, where they are commonly called Circle Cliffs or “racetracks.” Streams follow the curved valleys, which they leave eventually through a gorge.

The great fold that forms the Kaibab Plateau plunges down at both the north and south ends, so you would expect to see the curving pattern at each end. At the north end, the pattern is preserved and clearly visible from US 89. At the south end, the rocks have been eroded away, but the former curved pattern is revealed by the great bend of the Grand Canyon as it swings from a southerly to a northwesterly course around the nose of the Kaibab Plateau.
The Kaibab Plateau is higher than its surroundings today for one reason only: the Kaibab Limestone, which now is at the surface, is so much more resistant than overlying Mesozoic strata that these were completely eroded from the crest and flanks of the fold, whereas the Kaibab Limestone was essentially unaffected. The consequence is that, today, the Kaibab Limestone on the fold is high topographically, whereas the Mesozoic rocks away from the fold are low. On the other hand, streams flowing across the fold in the ancient curved valleys were stuck once they started sawing into the limestone: the sides of the valleys were completely eroded away, but the streams could not escape their limestone channel.
The concept for the carving of the Grand Canyon that I have been proposing for many years makes use of the elements of landscape evolution discussed above: the ancestral Colorado River flowed southwest toward the then-low-lying Kaibab Plateau, and crossed it along a racetrack valley near the Plateau’s south end. Once across the Plateau, the Colorado flowed northwest along one of the subdued valleys characteristic of the time and indicated to us by the lavas. The rocks into which this valley was carved were higher than present topography, so both rocks and valley have been eroded away.
We do not know where the river may have gone to once it left the Colorado Plateau. The topography of that country has been modified completely by the faulting that produced the modern basins and ranges of Nevada, and by volcanism that produced the Cascade Range. What’s more, the faulting has extended this region, making the modern coastline substantially farther away than it was when the ancestral river flowed toward it. Nevertheless, it is there that one needs to seek evidence by which to prove or disprove the concepts presented above, even if the changes that this region has endured make it unlikely that such evidence will ever be found.
The concept presented here provides a means for bringing the ancestral river across the Kaibab Plateau, which had been such an obstacle for previous researchers. The rest of the story is pretty much like that proposed long ago by McKee and colleagues: the Gulf of California opens, a new upstart river develops that starts worming its way into the Colorado Plateau, then taps into the ancient river, pirating its waters. In our case, however, the capture happens west of the Kaibab Plateau, not east. The old course of the river from the capture point to the west coast is abandoned suddenly and replaced by a new and shorter course through the western Grand Canyon to an arm of the Gulf of California only some 100 miles from the Plateau. The new river, steep of gradient and invigorated by the pirated waters of the old one, carves most of the Grand Canyon within the past 5.5 million years, and probably substantially less. The capture event is recorded in sand layers deposited by the Colorado River in the Imperial Valley of California, where fossils found only in rocks of the Colorado Plateau suddenly appear at a distinct level in the sand layers; none are found below that level.
Now what?
Rivers are shady characters that, like Chronus, devour their own children, the sand, gravel and cobbles they had once deposited. In doing so, they cover up their own tracks, making life difficult for the geologist bent on reconstructing the river’s history: there is plenty of slop and plenty of room for argument. Since it is not possible to prove geologic history the way it is possible to prove mathematical arguments, we geologists are forced to advance propositions, or hypotheses, as we call them, that we recognize as being provisional. However much we might dislike the idea, it is entirely possible that tomorrow someone might discover evidence that would force us to modify or even abandon our own beloved hypothesis. “Provisional,” however, is not the same as sloppy: the game has to be played according to well-defined rules. To begin with, a hypothesis must be based on observed facts, and must be in agreement with these facts. A hypothesis pulled out of thin air or—worse—not in agreement with known facts is not worth the paper it is written on. Second, a hypothesis that aims to supplant an earlier one cannot do so by simply ignoring it. What needs to be done is to show that the facts on which the earlier hypothesis is based are wrong, or that they were interpreted incorrectly, or that the new hypothesis is based on new facts or interprets the old ones in a more satisfactory manner. Since science is a collective effort that progresses through hypotheses, refinements, or refutations, ignoring what came before just doesn’t cut it.
Today, there has been an explosion of interest in the history of the Grand Canyon, and everyone seems to have a novel idea to set forth. Unfortunately, more than a few fail to conform to the scientific rules outlined above, and the proliferation makes it difficult to set down the main lines of thought in a lucid manner.
One of the most important arguments has to do with age of uplift of the Colorado Plateau. The reason is simple: no deep canyon can be carved in low-lying terrain, so the time when the Plateau was uplifted tells us when the Grand Canyon may have formed. Unfortunately, reliable evidence by which to time the uplift is hard to come by. Years ago I hoped to make a useful contribution to the problem by pointing out that we could use a five million-year deposit found along the course of the lower Colorado River. The deposit was universally regarded as marine or near-marine on the basis of its fossils, so it originated at or near sea level. Now, it is as high as 3000 feet, which can be taken as the approximate uplift of the region over the last five million years or so. So much young uplift agrees well with the youthful character of the Grand Canyon and with the information that there was no western Grand Canyon before about six million years ago, as explained earlier. Recently, however, people have argued that the deposit originated in salty lakes along the course of the lower Colorado, in which case it would not necessarily have been at sea level originally, and could not be used to determine uplift. This argument is based on isotopic data from the deposit, which indicate a fresh-water origin rather than marine. On the opposing side of the argument, fossils of many kinds uniformly indicate a marine or near-marine environment for the deposit, and the youthfulness of the canyons points to young uplift. The isotopic data themselves are more likely to reflect contamination than an actual environment. All things considered, I am of the opinion that a marine environment and young uplift remain the best interpretation.
Another notion that has gained attention recently is that the Grand Canyon was carved by the spilling over of a large lake. This Hopi Lake once occupied a good part of the Hopi Buttes country and was fed by the ancestral Colorado River flowing into it. The lake suddenly overflowed westward when its level rose above a topographic lip. This event realigned the ancient Colorado River into its present course and carved the Grand Canyon.
There are several problems with this idea. Those who have studied the lake most carefully believe it was not one large lake, but many small ones, and they also point out that a Colorado River emptying into these lakes over tens of millions of years would have filled them up entirely with sediment in short order. This issue can be attacked by studying the deposits of the lake to see if they contain material brought from the north by the ancient Colorado River. Another serious problem is that the present topography is used to identify a possible spillover point. But, as we saw earlier, the topographic surface was much higher even a short while ago than it is today, so any present-day lip has little application to past events. Even if water had spilled from the lake, its course westward from the spillover point could only be along a pre-existing drainage system, because land devoid of a drainage system is exceedingly rare. In other words, the spillover would have accentuated a pre-existing drainage system rather than creating a new one. Finally, ponding usually is an ephemeral event in a river’s history. Where did the ancestral Colorado go before the time of Hopi Lake?
Given these issues, the spillover hypothesis does not seem to fit in well with what we know about the ancient river system and the topography of the time, so I find it more reasonable to (provisionally) stand by the hypothesis presented earlier that the ancient river crossed the Kaibab Plateau when the Plateau was topographically low, then continued northwestward to a distant sea. This ill-defined continuation has been of concern to a number of researchers. The river would have flowed through that country now broken up by basin-range faulting. Are remnants of the old channel preserved on top of ranges there? Do the basins contain material brought in by the river? Was the region much wetter then than now, when the creation of the Sierra Nevada and the Cascade Range have shut off the supply of moisture from the west? Have ice-age floods and other events so modified the drainage pattern of the region that the old channels would not be recognizable even in areas not broken up by faulting? We simply don’t know. These are subjects well worthy of further study.
Finally, several people, starting with Charlie Hunt long ago, have suggested that underground water flow in Grand Canyon’s cavernous limestone layers such as the Redwall, may have been part of the river’s course in places. Such hypotheses are being tested by current work that is beginning to tell us when water flow through the caverns was taking place. Hunt’s idea that the ancestral Colorado River flowed south through Peach Springs Canyon has been disproved by more recent work.
By now you will have the feeling that pinning down the history of that old rogue, the Colorado River, is no easy thing: our ignorance is great and there is plenty of room for argument, which often is the more heated the less the evidence. What do I think? Well, it seems to me that my hypothesis still provides the best fit to the data on hand today, while being contradicted by none, so I’ll stick with it until someone shows that my data are wrong, or wrongly used, or comes up with new information that points to a better way. When and if this happens, I’d like to drink some good wine with the person who has made the discovery as a way to celebrate the progress made in working out this endlessly fascinating story.

Ivo Lucchitta

This is the seventh in a series of “Letters from Grand Canyon” by Ivo Lucchitta that will appear in future issues of the bqr. This particular “Letter” was divided into two parts.