Teddy and I were sitting about twenty yards apart. We had been like that for more than an hour, hunched up against the trunks of a couple of mopani trees as we waited for the herd of elephants to leave the grove we were in. They had moved into the grove from several directions and by the time we had noticed them we had lost any chance of retreating back to the Land Rover. There was nothing for us to do in the meantime but wait them out and to try to be as inconspicuous as possible. Climbing a tree was no refuge in this situation. That offers protection from cape buffalo, but not from elephant, which can reach the upper branches of trees with their trunks.
The elephants were not showing any indications of leaving anytime soon. The trees were full of succulent fruit and the elephants were munching away on them contentedly, pulling whole branches down for their young. Even under the trees it was swelteringly hot, and we were being tortured by tse-tse flies, which were biting us incessantly. We couldn't swat them away for fear of attracting the attention of the elephants and alarming them. Teddy had managed to squirm around to the far side of his tree, so that we could look out in both directions. We weren't worried that we would be trapped there until dark. There was no water in the grove and sooner or later the herd would head off to be near a water hole by late afternoon. Our big fear was that one of the small grazing groups would select one of our trees and be startled by our presence. There wasn't much we could do in that case. You can't outrun an elephant. One time a mother elephant came close enough to catch our scent in the still air, and snorting, wheeled away, whisking her calf in front of her with her long trunk.
We didn't have any water with us, and thirst was starting to claw at my throat. With dry lips I gazed desirously at some half-rotten fruit lying on the ground ten feet away, but I didn't dare try to go get them. I simply concentrated on watching the elephants and breathing smoothly, stifling any tendency to sneeze or cough.
After about two hours, trumpeting rang out from the south. The elephants near us perked up their ears and listened a moment, then headed off in a brisk trot in the direction of the trumpeting. They had been summoned for their afternoon trek to water: "Lunch break's over. Everybody back on the bus."
Finding ourselves alone at last in the grove, Teddy and I stood up and looked around warily. When we had convinced ourselves that there were no stragglers left, we started hiking back together to where we had left the Land Rover beside the track. "One thing I can say about you, Scholz," said Teddy. "You sure can pick the places to go to study earthquakes."
It was in the austral spring of 1974, and I had my reasons for being out there trying to record earthquakes in the hot sun of northern Botswana. In any scientific endeavor there are several types of reasons for taking on one particular project rather than another, just as there are different sets of goals. There are the goals of the customer, the agency or company that is paying for the work, and there are also genuine scientific goals, in which one seeks to solve an original problem which will contribute to the understanding of some subject. These may or may not coincide; that depends on the customer. If they do not, as in this case, the trick is to devise a research strategy that will accomplish both for the same price and effort, which usually means solving the basic scientific problem that addresses the customer's problem as a particular case. Whether research is applied or basic is often a matter of point of view and manner of execution. They are not necessarily mutually exclusive.
In some cases, and these are to be recommended, there are deeper-seated reasons for doing a project. These are more personal and intangible than the others. They might be called personal development goals: You take on a project because of the feeling, perhaps inchoate, that it may in some way contribute to your deeper understanding of the larger-scale research program you have chosen as your life's work. This type of motivation is hard to assess and even harder to evaluate later on in terms of its influence on your scientific development. It is just something that is felt, and doesn't bear thinking about too much. Pasteur said, famously, "Chance favors the prepared mind," and this is one way the mind is prepared. This type of feeling was a prime motivation for my taking on the Botswana earthquake project. Some of these hunches pay off, often in unexpected ways, and some don't, but they all add the kind of seasoning to life that a wild card does to a poker deck.
Not all scientists have such a well-defined vocation--that scientific subject to which they have decided to devote a lifetime of work. Many, perhaps most, scientists are simply people who have mastered a certain set of specialized skills and apply these to whatever work in which they may be profitably employed. In my case, however, I found my vocation early and have managed to find ways to pursue it, with few diversions, ever since.
I study the mechanics of the brittle deformation of the Earth. The outer fifteen to twenty kilometers of the Earth is composed of rock that is cold enough that when it deforms, it does so by brittle fracture, forming the shear cracks that geologists call faults. When further deformation causes the walls of the faults to slide past one another, this frictional sliding is often unstable and occurs in jerks, producing earthquakes. These two related phenomena, earthquakes and faulting, are the central topics of my research. My approach to them is nontraditional. According to the traditional way the earth sciences are organized, these two topics are the subjects of disparate disciplines: seismology, which purports to be the study of earthquakes, and structural geology, which encompasses the study of faults. Thirty years ago, when I went off to graduate school at M.I.T., I chose not to concentrate in either of those traditional fields, but instead to pursue a Ph.D. in the more arcane field of rock mechanics, the study of the mechanical behavior of rock. I chose this approach because the body of knowledge and skills taught in the traditional fields of seismology and structural geology provide the training only to describe their subjects, earthquakes and faulting, and not to explain how they work. I was not interested in the "when" and "where" questions, but in the "how" and the "why." This latter class of questions could be answered only if you first understood the basic material properties that govern the mechanical behavior. Thus, to my way of looking at the problem, faulting was a form of brittle fracture, and earthquakes a form of frictional instability, and both of these properties, fracture and friction, depend on properties of the material: the rock. The correct approach, in my opinion, was to study rock mechanics first, the other disciplines later. That was my starting premise and the way I arranged to be trained. Developments since have served only to strengthen this conviction. Modern theories of earthquakes and faulting are now all grounded in rock mechanics principles.
I work on fundamental problems of earthquakes and faulting, which are universal, rather than regional, in their application. Most of my work tends to be experimental or theoretical rather than field-oriented. When I use field observations, it is to check or demonstrate theory. If I study an earthquake in Japan, for example, it is because there is something in the observations of that earthquake critical to some theoretical point, and not because I am intrinsically interested in Japanese earthquakes. Although I am known in my field as an expert on Japanese earthquakes, this is mainly because Japan is the scene of multitudes of diverse types of earthquakes that Japanese scientists have been carefully studying for centuries, producing one of the world's greatest treasuries of earthquake data and lore. I have spent years sifting through this treasure-house of data and have written a number of papers on Japanese earthquakes and tectonics, but it would be more accurate to say that I am a student of Japanese datasets than of Japanese earthquakes, per se. I am a net user rather than a generator of field data.
I seldom do fieldwork myself nowadays, but things were not always like that. More than twenty years ago, when the events I describe here took place, I went through a period of several years in which I did quite a lot of fieldwork. I was under the impression at the time that I needed to become acquainted with my subject firsthand, in the field, to enhance my intuition of the subject and to properly appreciate the field observations made by others. Whether or not this is true is another story. I have never regretted my years doing fieldwork. That kind of work entails encountering trials and challenges of a qualitatively different cast than are found in other kinds of scientific work, and good fieldworkers develop their own special set of traits and skills to cope with them.
At the time that I took on the Botswana field project I was a freshly minted and tenured thirty-year-old associate professor at Columbia University, working at the Lamont-Doherty Earth Observatory in Palisades, New York, known more familiarly as Lamont. My career was pretty well on track, but I wasn't yet senior enough to have gotten bogged down in administrative duties. Back then, in the early '70s, federal support for scientific research was still in its heyday. A more free and easy attitude prevailed in selecting research topics than today, when the funding squeeze keeps most academic scientists in a tight caucus race of grant getting and paper production that precludes forays into risky serendipitous pursuits. So the prevailing conditions made it easy for me to take on the Botswana project, oddball though it had seemed at the beginning. It would be a stretch to say that it was very closely related to my main line of research. It was not something I had sought out; it had really fallen into my lap from out of the blue. But as I proceeded to investigate it, my curiosity was aroused, and I found it had greater depth than it had first appeared. The problem the customer needed solving was, from a seismological point of view, pedestrian, but it had a potentially important impact on a vital environmental problem. The scientific problem involved was, on the face of it, solely of regional interest, but there was a chance, a slight chance, that this little region of northern Botswana held the key to a much grander problem that has puzzled geologists for more than a century: How do the continents rift?
The customer was the U.N. Food and Agricultural Organization (F.A.O.), an outfit with which I had had no previous dealings. I was introduced to it by way of a phone call from an official at the U.N. Headquarters in New York who simply asked me straight off if I was interested in serving them as an earthquake consultant in Botswana. It was not the sort of question I was prepared to answer immediately, if for no other reason than because at the time I had only the vaguest idea of where Botswana was. I asked for more information, trying to winkle this out in the conversation, but he said he knew nothing about it. He was just calling me as requested by a cable from the F.A.O. headquarters in Rome. I replied that I might be interested, and to please ask the F.A.O. to send me a description of the consulting assignment, along with whatever background information that went with it.
Botswana? No doubt that was somewhere in Africa. Where, I wasn't too sure. My formative education in geography was from a wall map that had hung in my room as a boy, which as far as Africa went was from a decidedly earlier era. I remembered vividly the configuration of all the former colonies, like the Anglo-Egyptian Sudan and French Equatorial Africa, even down to their colors. The former was a pale green, the latter a Rand-McNally orangey-red, which was why I had had the notion for a long time that the former was grasslands and forest and the latter desert. I had had no particular reason to study Africa later on, so I still hadn't quite caught up with the postcolonial names and changes in borders.
I was soon put right by the Times Atlas. Botswana was the old British Protectorate of Bechuanaland. Independent since 1966, it is located north of South Africa and between Namibia and the country that was then, in 1973, still called Rhodesia. The bulk of Botswana consists of the Kalahari Desert. This partly explained my ignorance. My geographical knowledge had become specialized to seismically active areas. I had great familiarity with such earthquake-prone regions as the Kuriles and Kamchatka, New Hebrides and the Kermadecs, but to my knowledge the Kalahari was not among the seismically active areas of the world. The only seismically active region in Africa was the East African Rift System, which was, or so I thought, located far to the north of Botswana. The wall map I lived by at the time, "Seismicity of the Earth," confirmed this impression: It didn't show any seismicity in southern Africa. So why did the U.N. need an earthquake specialist in Botswana?
A packet arrived from Rome the following week. In it was a letter inviting me to become a consultant on earthquakes for the U.N. Development Programme on the Okavango Delta. Attached was a beautiful full-color brochure that described, in general terms, the Okavango program. It was full of photographs of an incredible place that looked like a lush tropical version of pine barrens. They showed forests of trees on sandy islands surrounded by papyrus-choked channels, and herds of elephant and antelope grazing near pools of improbably clear water in which hippos wallowed.
According to my atlas, which referred to this place as the Great Okavango Swamps, this was a huge area of freshwater swamp that lay, enigmatically, in the middle of the northern Kalahari. The brochure informed me that these swamps contained the only natural fresh surface water in Botswana, and were therefore considered a key to the future development of the country, particularly with respect to its abundant but as yet unexploited mineral resources. Various schemes were evidently being proposed for tapping these waters. At the same time, the vast migratory wildlife herds of the Kalahari depended on the Okavango for their sustenance during the long dry periods between the brief rainy seasons. There were, obviously, serious environmental conflicts in this situation, so the F.A.O. had undertaken this development program which involved making a broad ecological study of the region with the aim of assessing the impact of removing any portion of the Okavango waters.
So what did all this have to do with earthquakes? I read the letter again. It stipulated that in serving as an "earthquake consultant" that I would be expected to spend some time at the program headquarters in Maun, Botswana, and to prepare a report at the conclusion of my work. There was no hint at what the earthquake problem was or what, for that matter, I was expected to do once I arrived at Maun. I cabled Rome for clarification, but received back only an uninformative message stating that the need for the services of an "earthquake consultant" was called for in the master plan for the Okavango project, and that was why I had been contacted. It implied that I could expect to be filled in on my duties once I reached Botswana.
This was very strange. Hiring consultants is expensive, so a consulting assignment usually comes with a very well-defined charge. Here there seemed to be no charge at all. It reminded me of the instructions Lord Copper, the newspaper magnate in Evelyn Waugh's Scoop, gave to the neophyte reporter Boot on sending him off to cover a rumored war in Abyssinia: "Let me see. You will get there in about three weeks. I should spend a day or two looking around and getting the background. Then a good, full length dispatch which we can feature with your name." I was getting the impression that the U.N. people themselves didn't know what their earthquake problem was. It was almost as if they were saying, "Hey, you're the earthquake expert, you figure it out!" This was my first encounter with the sort of bureaucratic vacuity that was typical of the organization I was soon to be dealing with.
There was no way I was going to go traipsing off to Botswana on such a poorly defined mission, but my curiosity had been aroused. So I decided to do my own investigation of the tectonics in the region. I went downstairs to the computer center and loaded the tapes containing the locations of all sizable earthquakes that had occurred since 1962, when the Worldwide Standardized Seismographic Network had been installed. I instructed the computer to plot out a map of earthquakes in Africa. In those days plotters were mechanical devices in which pens, moving on mechanical arms, drew the plots on paper. After first making an outline of the continent, the plotter pen whizzed back and forth, pausing here and there and with a thwat-thwat making a cross locating the occurrence of an earthquake, the size of the cross indicating its magnitude. As the computer ran through its data file, the pen spent most of the time plotting crosses in the northeast part of the map, indicating earth quakes that had occurred along the East African Rift System.
With a length a sixth the circumference of the Earth, the East African Rift System is in the process of splitting the African continent into two parts. It is most prominent in the north, in Ethiopia, where it is opening the fastest. There the block containing Somalia is splitting away from the rest of Africa in an east-west direction. This split is marked by a series of deep rift valleys, containing many active volcanoes, that runs south from the Red Sea along the center of the broadly uplifted region that forms the Ethiopian highlands. Just south of Lake Turkana in northern Kenya the rift divides into two branches, an eastern one continuing to the south, passing to the west of Nairobi, and a western branch that follows the edge of the Congo Basin, forming the valleys filled by the great African lakes: Albert, Kivu, Edward, and Tanganyika. East of Lake Tanganyika the two branches rejoin and continue south through the rift valley occupied by Lake Malawi and the Shire Valley of Malawi, which are the last recognizable traces of active rifting.
The little crosses that were appearing on the map were accurately delineating this tectonically active feature, which is what I had expected to see. Occasionally, though, with a whirr the pen would transit far to the south and west and leave its telltale mark. So southern Africa was not as aseismic as I had thought. I made another map, this time just of the region south of Lake Tanganyika. On this map a faint pattern could be seen. There was, to the east, an irregular pattern of crosses extending from Tanzania through Malawi and petering out in northern Mozambique. This marked the southernmost and least active part of the rift system. To my surprise, I could also make out another, fainter line of epicenters farther to the west. This trend that I was now seeing was something new. Tearing off the computer plot, I overlaid it on a map of Africa and saw that this trend ran in a southwesterly direction from Lake Tanganyika, following the course of the Lwangwa Valley in northern Zambia, then continued along the course of the Zambezi gorges on the Rhodesia-Zambia border. There it was highlighted by a dense cloud of epicenters in the Kariba Gorge of the middle Zambezi. Following this trend with my eye, I found a few scattered events farther to the southwest, ending with a small cluster of four epicenters in the vicinity of the Okavango Delta. Bingo! Something indeed was going on down there. I was starting to feel the little tingle I get whenever I have spotted something interesting that nobody has noticed before.
I went into the archives where we keep our library of original paper seismograms and files of earthquake catalogs from the pre-computer era. Digging through the old catalogues, I found that two magnitude 6.5 earthquakes had been located in the Okavango in the early '50s. Unfortunately, very little was known about them. The coverage and quality of seismographic stations in those days would not permit a modern seismological study of such remote events, and, hardly surprisingly for that time and place, no contemporary field investigations had been mounted. Then I remembered that in the early '50s Lamont had installed a small global network of the first matched set of modern long-period seismometers. This had been the predecessor of a larger network set up for the International Geophysical Year in 1957-58, and which later evolved into the Worldwide Standardized Seismographic Network. In the part of the library that housed the records and data from that era, I found a map of this network showing that it had included three stations in Africa: Johannesburg, South Africa; Fort Lamy, Chad; and Helwan, Egypt. Entering the stacks, where in floor-to-ceiling racks thousands of 3-by-1-foot flat cardboard boxes of seismograms were filed according to station name and date, I was able to find the original paper seismograms for these stations. I pulled the boxes for June--August 1954, when the earthquakes had occurred. I spread out the seismograms for the appropriate days on the analyst's table, and found to my delight that both earthquakes had been clearly recorded at each of the stations. By measuring off the amplitude and arrival times of the various seismic waves at these stations, I was able to recalculate the locations and magnitudes of the earthquakes, using modern computer methods. These recalculations confirmed the figures given in the old catalogue, showing that the Okavango region was indeed capable of producing quite large earthquakes. The question was: Why?
I had noticed that the Okavango seismicity came at the end of a weak trend of seismic activity extending all the way from the East African Rift System east of Lake Tanganyika. Could this indicate a previously unrecognized branch of the rift system? The structure of the Lwangwa valley certainly resembled a poorly formed rift, but the continuity of this trend all the way to Botswana might be an illusion. The dense cluster of seismic activity in the Kariba Gorge I knew to be a man-made artifact. A high dam had been erected in the gorge during the '60s, and most of this seismic activity had been induced by the filling of its reservoir. The impoundment of large reservoirs has often been observed to induce seismic activity below them, as a result of the combined effects of the weight of the impounded water and of its hydraulic pressure in changing the state of stress underground. The induced activity at Kariba had been the subject of a number of detailed investigations using a local seismic network that had been specially installed for the purpose, and there was no doubt about the nature of those earthquakes. But now, armed with a new viewpoint, I could ask a different question of the data obtained by this local network. Was the reservoir-induced seismicity the result of triggering of small earthquakes in an otherwise inactive region, or was it an enhancement of activity in a naturally active belt? Going back to the archives, I found the monthly catalogues of the Rhodesian Seismic Network, which had been set up to study the Kariba seismicity. Close examination of the data given in those catalogues, with this question in mind, produced a definitive answer. I found that small earthquakes had been located in a trend along the gorge that extended as far as fifty to sixty kilometers downstream of the dam. The effects of the filling of the reservoir on earthquake activity would be negligible this far away from the reservoir, so these had to be natural earthquakes. There was then no doubt in my mind that there was, along the Zambezi gorges, a zone of natural seismic activity that simply had been stirred up by the filling of Lake Kariba.
The clincher for me arrived the following week, in the form of Landsat images that I had ordered from NASA. It is always spectacular to see these satellite images. Each one covers a region of about forty thousand square kilometers (200 x 200 km) and is printed in false colors diagnostic of moisture, vegetation, and rock and soil type. They show a level of detail greater than any map and over a scale range broader than can be visualized by any other means. To a geologist they are nothing short of wonderful: a treasure trove of information. They are called images, rather than photographs, because they are made using a number of spectral bands, several of which are not in the visual range, and because, having been gathered by several types of electronic instruments and constructed and reproduced by computer, they have little in common with what is produced with ordinary cameras or film.
The images were made at a time of the year, common for this region, in which there was no interrupting cloud cover. They showed an amazingly flat region, devoid of almost any topographic feature except the occasional blemish of a kopje, the isolated little rocky hills of southern Africa that poke out of the desert like islands in a sea of sand. Along the northern edge of the region the Zambezi flowed from west to east, but had not yet entered, at Victoria Falls, its series of deep gorges, and so flowed, desultorily, in a meandering series of swamps and braided streams. In the northwest corner of Botswana, in the area bordering Angola and the narrow salient of Namibia called the Caprivi Strip, there was a vast empty region covered by long linear sand dunes running east and west. Through this meandered, in a generally southeastern direction, the Okavango River, which drains the well-watered Angola highlands. Then suddenly, in this almost totally flat terrain, the river split into hundreds of channels, which diverging, formed an ever expanding filigree of anastomosing streams and pools. It looked as though an enormous glass of water had been spilled over the desert, and every point that had been touched by the water had turned bright green.
More astonishing still, a hundred miles or so downstream from this point of divergence, all of these myriad streams again rejoined, collecting into a river perpendicular to the first, and forming this whole watery region into a roughly equilateral triangle, some 120 miles on a side. This was the Okavango Delta. I could also see why this was a freshwater delta. It had two outlets, one to the south, which terminated in the saline Lake Ngami, and one that continued farther to the southeast, finally ending in the great Makgadikgadi Salt Pans.
What a curious situation this was. For very good reasons, rivers do not generally branch in a downstream direction. And, under the combined effects of gravity and erosion, they usually flow directly to the lowest elevation in the region, where, if there is no outlet, they evaporate to form a salt flat or saline lake. They do not usually form inland freshwater deltas like this, perched above their natural hydrological declivities. And even if they do, they have one, and not two, outlets. This was not a stable situation. There was some other agent, beside the normal action of flowing water, involved in producing and maintaining this peculiar arrangement.
I pored over all the images. I had taken the precaution of ordering a full set of images at each individual spectraI band, and in several combinations, since the different spectral frequencies are sensitive to different kinds of features. I first focused on the Thamalakane River, which formed the downstream base of the triangle of the Okavango, because it was suspiciously straight along much of its course. Looking carefully, I could also see subtle linear features running from southwest to northeast across a number of the images. Two of these ran from Lake Ngami across the base of the delta, between which flowed the Thamalakane, the river that collected all the delta streams. Parallel lineaments, offset in an echelon manner continued through the veld of the Mababe and Chobe regions to the northeast as far as the border with Rhodesia and Zambia. Another, similar set of features could be seen at the head of the delta. These were all parallel to the trend of the line of seismic activity that passed through the Kariba Gorge and continued on to the Okavango. Bingo, and bingo again!
There are very few features in nature that are really straight, and at this scale, tens to hundreds of kilometers, there is only one likely suspect: a geological fault. If these lineaments were active tectonic faults, extending as it seemed in a trend from the basin of Lake Ngami all the way to the Zambezi gorges, then the most likely explanation was that I was indeed looking at the tip of a previously unrecognized branch of the African rift system. That would explain the seismic activity, and more. If these faults had the same sense of motion as the faults that bound the African rifts, then the Okavango Delta was located within a nascent, now barely perceptible, down-dropped rift valley. This one was about a hundred miles wide somewhat wider than the great rift valleys of East Africa; unlike those, the bounding faults in this one must have just started to move. Here there were no great escarpments dropping the valley floor a kilometer or so below the level of the surrounding plateaus. Whatever depression had occurred by fault motion was masked by being infilled with sand. At the head of the delta the river crossed several faults, which, I guessed, produced a sudden increase in slope in the downstream direction, contrary to the normal hydrographic profile of rivers, so the stream split there into many branches. At the distal end of the delta the waters encountered another fault which was slipping the other direction, with the northwest side moving down. That created a low scarp facing upstream, which acted as a natural dam, channeling the Thamalakane and empounding the Okavango water above the local minima in elevation: Lake Ngami and the salt pans. To maintain this arrangement these faults had to be actively moving to counter the normal erosional action of the rivers, which, if given enough time, would otherwise destroy these unstable features.
The hypothesis had the delightful feature of explaining things on two very different scales. Regionally, it could account for the presence of that faint band of seismicity as delineating a previously unrecognized arm of the African rifts that extended all the way from Tanzania to somewhere just south of the Okavango. On the local scale, it also explained why the Okavango existed. This latter point tickled my sense of whimsy. Seismologists have known for some time about reservoir-induced earthquakes, as at Kariba, but here was the first recognized case of an earthquake-induced reservoir!
I was now beginning to realize that there was an intelligence behind this assignment, after all. Any earthquake occurring on one of these faults would produce fault motions that were mainly vertical. An earthquake as large as the ones of the '50s could be expected to produce a few meters of uplift and subsidence, which in such a flat area would be enough to seriously alter the drainage pattern within the Okavango. The whole hydrological system would be put out of plumb. Any development scheme that would seek to tap these waters would have to accept the chance that it could be seriously disrupted by an earthquake of that size. Because the whole region was tectonically active, the topography was unstable and the stream pattern was therefore intrinsically ephemeral. The customer did have a potentially major earthquake problem, after all. What it needed to know was how likely this possiblity was, and how serious its consequences.
Return to Book Description
File created: 8/7/2007
Questions and comments to: email@example.com
Princeton University Press