Titicaca: An Archive of South American Paleoclimate
Paul A. Baker, Sherilyn C. Fritz and Geoffrey O. Seltzer
Climate can be simply defined
as average weather, and paleoclimate as the extension of climate further back
in time. The most important climatic and paleoclimatic elements, in terms of
impact on Earth surface and life, are temperature and precipitation. How temperature
and precipitation vary in space and time is thus the main focus of climatology
One of the biggest differences between the two disciplines is how these variables are measured. Another important difference is that paleoclimatologists cannot begin to fathom how to directly determine past atmospheric pressure gradients, the major control on atmospheric circulation. Whereas climatologists utilize meteorological instruments on weather stations, ships, moorings, balloons and satellites, the main tools of paleoclimatologists are corers, microscopes and mass spectrometers. The archives of paleoclimatic history include glacial ice, tree rings, corals and sediments deposited in the oceans, rivers and lakes.
Each of these different types of paleoclimate recorders has its own advantages and disadvantages. For example, corals lay down their calcium carbonate skeletons fast enough to record an annual or even shorter temporal resolution. But corals are somewhat imperfect recorders of climate (specifically, of sea-surface temperature). Also, reef-building corals only occur in shallow tropical oceans, and it is difficult to obtain specimens that span time periods greater than the past few hundred years.
Lakes are distributed over much of Earth's surface. Lacustrine sediments contain a rich variety of constituents - fossil diatoms and ostracodes, pollen, organic detritus, volcanic ash, fluvial and glacial detritus, and more. These constituents can serve as valuable paleoclimatic proxies. But these proxies are always indirect - there is no perfectly accurate paleo-rain gauge or paleo-thermometer. Lacustrine sediments generally accumulate at a rate of a millimeter per year, about 100 times faster than most marine sediments. Thus the temporal resolution of lacustrine records is generally much better than for marine sediments, but does not usually approach the resolution of coral or tree ring archives. Lacustrine sediments often span many thousands and, rarely, millions of years.
Recovering these records often brings paleoclimatologists face-to-face with difficult challenges and exciting adventures. We have had both during our ongoing efforts to uncover the paleoclimatic history of tropical South America by coring Lake Titicaca. One of our primary motivations for studying Lake Titicaca was that, because it is the largest lake in South America by volume, we presumed it held a long and continuous record of the paleoclimatic history of this part of South America and the neighboring Amazon basin.
Lake Titicaca is known to many people as the highest lake in the world, which it is not - that distinction is now said to belong to a small crater lake situated atop a lofty volcanic peak, Licancabur, some 500 kilometers further south. But Lake Titicaca is high (3,810 meters above sea level), large (8,600 square kilometers), deep (285 meters maximum), and cold. It is also a tropical lake located between about 15 degrees and 17 degrees south latitude.
Lake Titicaca is in the northern portion of the Altiplano (Fig. 1), a large, internally drained high altitude plateau sandwiched between the eastern and western cordilleras of the Andes. Although the lowest point on the surface of the Altiplano is actually the floor of Lake Titicaca, when the lake is full, as it is in modern times, it drains southward via the Rio Desaguadero into shallow, saline Lago Poopo. The climate becomes increasingly arid as one travels southward on the Altiplano. Annual precipitation decreases from 1,000 millimeters at Lake Titicaca, to 400 millimeters at the city of Oruro near Lago Poopo, to less than 100 millimeters in the southernmost Altiplano just east of Volcan Licancabur.
Despite its tropical latitude, Lake Titicaca's height keeps its waters cold. The large expanse of the lake surface provides adequate fetch for development of large waves during windy periods. The relative lack of modern commercial interests on the lake ensures little boat traffic. Taken together, these factors contributed to our first challenge: how to safely navigate and extract sediments from the open-water, deep portions of the lake.
In 1997 we obtained the research vessel Neecho from the U.S. Geological Survey. With a grant from the National Science Foundation and a lot of work by James Broda of the Woods Hole Oceanographic Institution, the Neecho was refurbished, then shipped from Massachusetts, through the Panama Canal and to Arica, Chile. From Arica, a tractor-trailer ferried the Neecho over the western cordillera of the Andes to its new home port of Huatajata, Bolivia.
During our maiden cruise on the Neecho in May 1998 and on several subsequent expeditions, we completed many hydrocasts, conductivity-temperature-depth casts, plankton tows, and seismic-reflection profiles, all of which allowed us to learn about the physical, chemical, biological and geological processes in the modern lake.
The bread and butter of our paleoclimatic studies are the sediments raised from the lake floor in piston cores and box cores.
A day on the lake
On a typical coring expedition
the three of us, accompanied by one or two of our graduate students, assemble
in Huatajata in May or June (the Austral fall) with our mechanic, Gonzalo Mollericon,
and our pilot, Nicolas Katari. Nico and Gonzalo are both Aymaran, the majority
culture around the Bolivian and Peruvian shores of Lake Titicaca. Their intelligence,
knowledge, resourcefulness and political skills have brought us through many
difficult experiences. Together, we overhaul the Neecho after its months at
mooring, load supplies, and top off the fuel tanks. We set out with twin diesel
engines straining to make 10 knots in the rarified air, and arrive an hour later
at the narrow Straits of Tiquina.
The Straits frame our entry into the larger and deeper Lago Grande and our exit from the quiescent waters of Lago Huinaimarca. To our right are the heavily glaciated peaks of the Cordillera Real, rising another 2,500 meters above the lake. On our left are the heavily terraced slopes of the Tiquina peninsula and Isla del Sol. The terraces are relics of the farming methods of the ancient cultures of the Tiwanaku and the Inca, and reminders of the importance of this region to those people. Ahead of us, looking northwestward, we see nothing but water.
On earlier expeditions we learned that as recently as 5,000 years ago, the surface of Lake Titicaca had fallen 100 meters below its present, overflowing level. Prior to 5,000 years before present, modern lake floor shallower than 100 meters would have been exposed and subject to erosion. Thus, if we want to recover a core that contains a continuous sedimentary record, we must make for deeper, open waters. To discover the nature and structure of the bottom sediments, to determine the depth of the lake, and to identify the best coring sites, we continuously record seismic reflection data. These data may show continuous parallel reflectors (generally good places to core) or they may reveal complex folding, faulting or unconformities, which themselves provide information about the evolution of the lake basin and its filling history, but are less desirable locations for our coring.
On the Neecho, piston cores are taken just as they are on larger oceanographic vessels. The corer is a series of 10-foot sections of stainless steel pipe. These are bolted together in a horizontal position along the side of the vessel. We attach a temporary catwalk to the bow so that we can assemble a core that is longer than the Neecho's 13 meters. A clear plastic liner is inserted inside the core pipe. We run a wire rope that is three-eighths of an inch thick from a large hydraulic winch attached to the deck and through a block on an A-frame mounted on the stern. Then we take up a couple of loops of slack, attach the rope to a trigger assembly, run it down the inside of the core pipe, and attach a piston to the end of the rope. The piston is then inserted into the forward (but soon-to-be bottom) end of the core pipe. We then swing the core pipe into the vertical position and rotate astern beneath the A-frame. We use the winch to lower the corer to a few meters above the lake floor. The corer is triggered and free falls (taking up those extra loops of slack) into the soft bottom, driven downward by its own considerable weight. We then reverse the procedure: the sediment-filled liners are removed, capped and labeled. The real work begins.
The climate stories
These cores are the raw materials
for our paleoclimatic studies. In Lake Titicaca, our longest piston core contained
13.5 meters of sediment spanning the past 35,000 years. The cores are continuously
logged for magnetic susceptibility and described using a variety of traditional
sedimentological methods: visual description of lithology and sedimentary structures,
petrographic microscopy, physical properties, grain size, etc. Other measurements
that we or our co-investigators make include diatom and pollen assemblages,
stable isotopic compositions of calcium carbonate and organic matter, and sediment
The Last Glacial Maximum (LGM) occurred worldwide about 21,000 years ago and was marked by a 120-meter drop in sea level, an expansion of continental ice sheets in northern high latitudes, and a decrease in global temperature of some 5 degrees Celsius. The nature of the LGM in tropical South America has been elusive, with most investigators positing greatly increased aridity throughout the Amazon basin, indeed throughout the tropics worldwide. This hypothesis was a logical deduction from thermodynamics: If the tropical oceans were colder during the LGM, then the atmosphere above them and land masses adjacent to them must have contained less moisture. The argument, however, does not take into account changing atmospheric dynamics, such as possible increased transport of moisture from the oceans to adjacent continents by intensified trade winds. Nor does the argument consider the changing patterns of moisture from changing seasonality brought on by Earth's orbital variations (discussed in more detail later). Furthermore, at least in the case of tropical South America, the argument was based on almost no direct paleoclimatic evidence.
Investigating our sediment cores, we observed that 21,000-year-old sediments are full of glacial detritus, but that younger sediments lack glacial detritus. The contrast strengthens the notion that glaciers expanded in the tropical Andes surrounding the Altiplano coincident with the global LGM. This regional glacial expansion was partly due to the global cooling, but was also due in part to increased precipitation. As a result of increased precipitation and decreased temperature - hence, decreased evaporation - Lake Titicaca was a fresh and overflowing lake throughout the LGM (Fig. 2).
Overflow spilled down the Rio Desaguadero, filling Lago Poopo and the salars and creating a giant paleolake (drill cores that we and Catherine Rigsby from East Carolina University took in 1999 from the now-dry Salar de Uyuni contained deep lacustrine sediments spanning the LGM). This evidence points toward increased moisture, not increased aridity, throughout southern tropical South America and most of Amazonia during the LGM.
In the subtropics, solar radiation varies greatly throughout the year, generally producing a summer wet season and a winter dry season. The amplitude of the annual cycle of solar radiation itself varies on long timescales with periodic variations of Earth's orbit. The period with largest amplitude is 23,000 years. Thus, every 23,000 years, Earth experiences a maximum in solar radiation during the summertime, followed 11,500 years later by a minimum, and then another maximum.
The amplitude of this "precessional" cycle is about one-third of the annual cycle of solar radiation, so it is undoubtedly significant in forcing an increased or decreased South American summer monsoon, as many call it. During the LGM the summertime radiation in the southern tropics was near a maximum. We believe that this is the major reason why the southern tropics witnessed a precipitation maximum at the same time. Roughly 11,500 years later (10,000 years ago), summertime radiation and precipitation reached minima and the level of Lake Titicaca dropped by 100 meters or so. By 3,500 years ago, summertime radiation and precipitation again increased enough that the lake attained nearly its present level.
Large changes of precipitation and lake level evidently occur on shorter timescales as well (Fig. 2). Between 8,000 and 9,500 years ago, almost precisely at the time of minimum summertime insolation, the lake level unexpectedly rose and once again began to overflow southward. Prior to our coring, this highstand was unknown and unsuspected. Its origin is vexing, but we believe that it is related to distant atmospheric anomalies identified in studies of Greenland ice cores and related to oceanographic anomalies that were identified in sediment cores from the North Atlantic. We have hypothesized - but our hypothesis remains controversial - a general linkage, at timescales ranging from decadal to multi-millennial, between cold sea-surface temperatures from the North Atlantic and increased precipitation on the Altiplano and probably throughout much of Amazonia. Although such a relationship has not been clearly identified in modern instrumental measurements of climate, it is physically plausible and suggests avenues of study for climatologists.
In April and May of 2001 - with funding from the National Science Foundation and the International Continental Scientific Drilling Program, and with technical expertise provided by DOSECC Inc. - we initiated an exciting project to drill deeper into sediments on the floor of Lake Titicaca in order to recover a longer paleoclimate record. We recovered a total of 625 meters of sediment at three sites. The longest single drill core recovered 136 meters of sediment in 235 meters water depth.
Our studies of these sediments are only beginning, but in years to come we hope to uncover a great deal more about the nature of past climate change and the mechanisms effecting this climate variability.