The Dry Valleys of Antarctica: Data for Big Science

By Jane Beitler, NSIDC

A wind-eroded rock sits in front of Canada Glacier in Antarctica’s Dry Valleys region. Photo courtesy Chris Gardner/2006 McMurdo Dry Valleys LTER


In Antarctica lies a landscape that looks more like Mars than any place on Earth. Deep, barren valleys crisscross the Dry Valleys region, west of McMurdo Sound, absent the snow and ice cover that are typical of the continent.


Yet signs remain of molten-hot volcanic processes that formed the Earth and the Dry Valleys landscape. Erosion has carved through layers of rock to a depth of up to four kilometers, over a ten thousand square kilometer area. To petrologist Bruce Marsh, at Johns Hopkins University, it is a window into the past. Marsh said, “This is the most unusual part of the world. It’s some of the oldest terrain on Earth, and nothing has happened there since 60 to 80 million years ago, when it was eroded deeply, like Monument Valley or the Grand Canyon. It’s like going into an ancient Egyptian library and finding it all intact.”


Not even a lifetime, let alone the nineteen years Marsh has been studying the Dry Valleys, would be sufficient to unlock all their secrets. So he is working with the Data Conservancy to make his research data accessible to others, now and in the future so that others can build on his observations. In a new twist, the data are already helping scientists who study glaciers and climate change.


Ancient mysteries revealed


Data scientist Keith Kaneda, at the Sheridan Libraries at Johns Hopkins University, was looking for a test case when he contacted Marsh. As part of the Data Conservancy, a National Science Foundation-sponsored collaboration, Kaneda and his colleagues are working on a blueprint for managing science data, which today’s science generates at an increasingly fast pace. The needs for science data are analogous to what libraries provide for text-based knowledge: public places to store, preserve, find, and get data in a wide variety of forms. Kaneda needed a rich case that exemplified the complexities and promise of this challenge.


Kaneda said, “I started talking with Bruce about his research, and my jaw dropped. What he has been studying in the Dry Valleys is a fossil magmatic plumbing system, perhaps the best-preserved and exposed example of its kind on Earth. One of the long-standing geologic mysteries is an understanding of the integrated processes acting on magmas as they ascend from their source regions, through their magmatic plumbing system, to erupt at the Earth’s surface. Although bits and pieces of these plumbing systems occur worldwide, petrologists have never had a complete example to study.” Understanding these magmatic systems is central to learning how the Earth and its igneous rocks were formed, and continue to form.


The photograph above shows exposed layers in the Dry Valleys, visible to a depth of approximately four kilometers. Here, four horizontal sheets or sills of igneous rock are exposed between layers of granite and sandstone; the sills were formed by magmatic intrusions 180 million  years ago when the ancient super-continent of Gondwana broke up, and the exposed valley walls provide insight into ancient magmatic flows (show in grey and green in the diagram at right) deep below the surface. The top of the system, where volcanic cones once may have formed, is apparent on the horizon. (Courtesy Bruce Marsh, JHU)


Early Antarctic explorers had reported exposed layers of fossilized magma in the Dry Valleys, but until recently, petrologists had not studied the valleys in detail. The layers are remnants from the processes that formed Earth’s continents. “I went down there originally to look at how the Earth evolved,” Marsh said. “I heard the system was a mess and didn’t make any sense. But by the second season there, I began to think it had an order far beyond anything we ever realized. As we got into understanding the overall structure, I was overwhelmed. It was more detailed and more ordered and more understandable.”


Kaneda said, “What’s significant about the Dry Valleys is the exceptional degree of preservation and exposure of the whole magmatic system. It is a natural laboratory for studying and understanding magmatic systems and processes, perhaps the finest that petrologists have ever had an opportunity to study.” Marsh added, “It’s like going back on Earth three billion years ago.”


The igneous rock exposed in the Dry Valleys dates back to when the continents broke up in the Jurassic period. “There are big magmatic sheets, a whole stack of them,” Marsh said. By examining the crystalline structure of the exposed rock, Marsh traced the processes of ancient eruptions. “For example, some of the erupted bodies were so hot, they started to re-melt the crust,” he said.


Scientists are far from having the whole picture of the processes that created today’s continents. Kaneda saw Marsh’s work as a good example of data that could support the mind-bogglingly immense science questions that require access to much data, often spanning long time periods, multiple research disciplines, and formats from satellite to field observations.


Yet even when relevant data exist on a particular topic, accessing the data may be difficult to impossible. The growth of science data has outstripped the infrastructure and practices to preserve the data and make it accessible. Data may reside in isolated repositories, where it can be hard to discover. Many data sets, like the observations from Marsh’s years of field study, are not accessible at all. This is just the set of problems that Data Conservancy was charged to help solve.


More than just numbers


Marsh has amassed thousands of objects and observations during his summers in the Dry Valleys. Kaneda said, “People think of data as numbers, but in a geologic field study there are many different types of data.” Kaneda is working with Marsh to describe and catalog types such as geochemical analyses, rock samples, thin sections of rock, and photographs. Kaneda said, “The goal is to preserve the data and their context so that a user of the data will come as close as possible to having access to the same information that Bruce Marsh and the members of his research group do.”


Preserving data and their context is no simple task when working with disparate types of data. In some cases, there is no clear model of how to do this. Kaneda said, “A thin section is a slice of rock that is ground thin enough for light to pass through it. You use a petrographic microscope to examine the rock to look for evidence of where and how the rock formed and processes that affected it. With physical objects like this, how do you preserve the information and make it available to other researchers?”


Kaneda is hoping to borrow a new idea from the medical and biological sciences: virtual microscopes. This technique puts high-resolution images of thin sections on the Web.


Marsh also took many photographs of the valleys and their exposed layers. Kaneda asked him to annotate the photographs so that he could create metadata, a structured system of describing the data. During this process, Kaneda’s colleague Ruth Duerr realized the photographs also contained information that was useless to Marsh, but of great interest to other scientists. Marsh had inadvertently documented the conditions of glaciers in the valleys, which have been changing rapidly over the last decade and a half. Today, the Data Conservancy has made those photographs and their metadata available through the National Snow and Ice Data Center, as part of their large collection of glacier photographs.


Petrologist Bruce Marsh studies exposed relict terrain in McMurdo’s Dry Valleys to better understand the magmatic processes that formed Earth’s continents. While photographing this terrain, he also captured images of Antarctic glaciers, such as this one of Goodspeed Glacier (left), Hart Glacier (center), and Meserve Glacier (right). His photographs help document the state of these glaciers, which have been changing rapidly in recent decades. (Courtesy Bruce Marsh/NSIDC Glacier Photograph Collection)


Data for the future


Re-use of data by different disciplines is only one goal of the Data Conservancy. Equally important is the need to preserve data for future use. Kaneda said, “As a researcher winds their career down and retires, often their research materials—notebooks, data—do not get preserved. Materials get lost or thrown out because they don’t have specific staff or facilities to properly curate them.” Data scientists like Kaneda see this as a waste of potential. “This is the life work of these people. It represents years of time and effort, as well as public money spent, and if the results have not been preserved, it is a lost investment.”


Kaneda and his Data Conservancy colleagues are busy developing the means to preserve all types of science data. Working with actual cases like the Dry Valleys data helps them ensure that systems can accommodate different data types and sciences. He said, “There are many groups working hard on different aspects of data curation. One group is building software and hardware infrastructure. For the Dry Valleys project, we are continuing to digitize the data, and to create and associate the metadata with Marsh’s data.”


The metadata preserves the context for the data that is essential to understanding and using it. For example, it is important to know exactly where a rock sample was collected. “The Data Conservancy is not only trying to preserve the data; we want future researchers to be able to use the data,” Kaneda said. As the infrastructure develops, Data Conservancy managers plan to add even more sophisticated access tools for researchers. Kaneda said, “We would like to eventually to build some sort of application with a geospatial interface. That will allow people to access the data and to see the different types of data in context, and get an accurate picture of the Dry Valleys area.”


Increasingly, science is based on data. The Data Conservancy’s work helps ensure that data becomes part of the repository of scientific knowledge, opening possibilities for knowledge to grow in leaps and bounds. A researcher halfway across the world or years in the future may gain new insights by discovering and incorporating Marsh’s data into a new study. Kaneda said, “You can think of this like trying to democratize access to physical objects. It’s similar in some ways to books: we want to make them available to people in places that don’t have good libraries.”


Building the future of knowledge


In the process of documenting his data, Marsh is getting a fresh look at his own data from years ago. “I thought it would be a straightforward recording,” Marsh said. But he got unexpected benefits from the effort. “By going through and digitizing my slides, it actually helps me in my present research, by re-familiarizing me with the data,” Marsh said. Marsh has also used the Data Conservancy search tool to quickly locate what he needs in his own collection.


Marsh finds it gratifying to help preserve his data. He said, “Today, people recreate the wheel. They go to the field and they have seen my published paper, but they have to go collect their own samples because they don’t know where Marsh’s samples were collected from, what the geologic context was where he collected them, what they look like, in what conditions he collected them, and so on. With things like thin sections and pictures of the rocks in the field where we were taking samples, people in the future can search for these things. They can pick up right where I left off.”


Kaneda and the Data Conservancy plan to make these data curation services and systems a standard fixture of research and academic infrastructure for the sciences, just as today libraries are essential repositories for written knowledge in the humanities. Marsh said, “I find it very exciting that I am the first guinea pig. We are in the transition stage into the digital world, where all data will be available to everyone: in the field, in real time, in Antarctica. We’re like new people on a new planet: a digital planet.”


Marsh’s research party stands in the Simmons Valley of Antarctica’s Dry Valleys, with Taylor Glacier in the background. When the team took this and similar photographs for their petrologic research, they did not realize that other scientists would be interested in the glacier images that they captured. (Courtesy Bruce Marsh/NSIDC Glacier Photograph Collection)


For more information:


The Data Conservancy:


Bruce Marsh—Research Web site:


National Snow and Ice Data Center—Glacier Photograph Collection:




Kaneda, Keith, et al. 2010. Data Conservancy: Digital curation of the magmatic system of McMurdo Dry Valleys, Antarctica. Presented at the 2010 Geological Society of America Denver Annual Meeting, 31 October – 3 November 2010.


Marsh, B.D. 2004. A magmatic mush column Rosetta Stone: The McMurdo Dry Valleys of Antarctica. EOS Transactions 85, no. 47, 23 November 2004, 497-502.


Heyn, J., Marsh, B.D., and Wheelock, M.M. 1995. Crystal size and cooling time in the Peneplain sill, Dry Valley Region. Antarctic Journal of the United States 30, 50-51.


Marsh B. D. and Wheelock M. M. 1994. The vertical variation of composition in the Peneplain sill and Basement sills of the Dry Valleys: The Null Hypothesis. Antarctic Journal of the United States 29, no. 5, 25-26.

Posted in Uncategorized | Tagged , , , | Leave a comment

Leave a Reply

You must be logged in to post a comment.