Scientists reconstruct collapsed Antarctic glaciers using 1960s aerial photographs

Antarctica’s Larsen Ice Shelf has been breaking up for decades, but the 2002 Larsen B collapse was particularly dramatic. After being stable for at least 10,000 years, a large section of the shelf broke apart, with repercussions felt across the planet.

The widespread changes in Antarctica have been extensively studied and published, but contextualizing and analyzing how the changing conditions in Antarctica affect the rest of the world has proven challenging. To combat this, scientists have used film photographs from the 1960s to help understand how collapsing Antarctic glaciers have affected global sea levels.

In a research article published in NatureRyan North and Timothy T. Barrows examine historical film photographs of Antarctica dating back to the 1940s and apply an advanced modern analytical technique called structure-from-motion (SfM) photogrammetry. Researchers say, “The technique creates digital elevation models (DEMs) by constructing 3D point clouds of corresponding features in overlapping photographs without requiring the original camera positions or orientations.”

A satellite image of a largely snow-covered landmass with visible dark waters and scattered icebergs. The left side has darker patches of water alternating with white ice formations, while thick clouds dominate the right side of the image.
The collapse of the Larsen B Ice Shelf, seen from space on March 17, 2022. | NASA

As North and Barrows explain in an article on The conversationAn accurate understanding of the past is essential to predicting the future.

“To accurately predict how Antarctica’s glaciers will respond to future climate change, it is critical to understand how they have responded in the past,” the researchers write.

Composite image showing different perspectives of Antarctic glaciers around the Larsen B Ice Shelf. Different maps in panels show elevation changes, ground contours and glacier positions, with emphasis on areas such as Melville, Crane, Mapple, Jorum and Flask Glaciers.
Figure 1. This shows an example of overlapping historical aerial photographs from December 1968, the resulting three-dimensional models derived from the historical aerial photographs, and associated glacier contours. The ice shelf contours are from the SCAR Antarctic Digital Database.

A major challenge is that Antarctica is remote and capturing good data there is prohibitively expensive. While satellites are excellent at collecting data over a large portion of the Earth’s surface, the Antarctic Peninsula is shrouded in cloud for much of the year. As a result, observations of the area are patchy and short-lived.

However, U.S. Navy cartographers took more than 300,000 aerial photographs of Antarctica, all of which are available free of charge from the University of Minnesota’s Polar Geospatial Center. They were part of a large-scale mapping project of the continent from 1946 to 2000.

The large-format film photographs have extremely high resolution. Therefore, North and Barrows applied SfM photogrammetry to 871 specific photographs from 1968 to construct historical data for the Larsen B region.

The selected photographs were captured on large format 9×9 grayscale film on December 21, 23, and 27, 1968. The film was then scanned at 1,000 dpi by the United States Geological Survey (USGS). 503 of the 867 photographs were used to construct elevation data for the Jorum, Crane, Mapple, and Melville Glaciers, while over 360 were used to determine the elevation of the Flask Glacier. The images were also manually refined to reduce errors in the photogrammetry process, including changes in cropping, exposure, contrast, and brightness.

“We use the historical DEMs to precisely measure the net change in surface elevations of Larsen B side glaciers (Jorum, Crane, Mapple, Melville and Flask) between 1968 and 2021. For the same glaciers, we also calculate the differences in surface elevation between 1968 and 2001… Using precise elevation changes, we provide new estimates of the mass balance and sea level contribution over a 53-year period and discuss these measurements in the context of existing literature covering the pre- and post-collapse period,” they explain.

A four-panel image shows satellite images of Crane Glacier, with highlighted boxes for areas of interest. Each panel zooms in closer, culminating in the detailed view of surface debris and a meltwater stream in the final panel.
Figure 2. ‘Details of a historical high spatial resolution orthophoto mosaic (pixel size 1.6 meters) covering the Larsen B area in December 1968. (A) The full extent of the orthophoto mosaic, (B) a magnified area on Crane Glacier, (C) a tributary of Crane Glacier that has been further magnified, and (D) meter-sized surface debris and meltwater channels visible in the same tributary of Crane Glacier.’

As a result of their research, the duo determined precise elevation changes in individual regions of Larsen B, detailing the exceptionally small changes over decades that ultimately led to the ice shelf’s collapse. They found that after Larsen B collapsed in 2002, the glaciers lost a staggering 35 billion tons of land-based ice. The largest glacier alone lost 28 billion tons, resulting in a sea-level rise of about 0.1 millimeter.

“That doesn’t sound like much,” the researchers admit. “But it’s the result of one glacier from one event. In other words, it’s the equivalent of every person on Earth chugging down a one-liter bottle of water every day for 10 years.”

Map of changes in surface elevation in glaciers from 1968-2021 in the Antarctic Peninsula. Areas in red show significant elevation loss, particularly in Mappie Glacier. Larsen B Inlet, Crane Glacier, and other landmarks are highlighted.
Figure 5. This shows the elevation change of Larsen B and its associated glaciers from 1968 to 2001. The net mass balance is indicated as the mouth of each tributary.

North and Barrow call the historical archive of aerial photographs an “invaluable resource waiting to be tapped” and say the same process they used for this study could be used to analyze other ice shelves or glaciers, penguin colonies, vegetation expansion, or even direct human activity.

Antarctic ice shelves and glaciers will continue to advance as climate change accelerates, with consequences for the rest of the planet. One of the most crucial steps to addressing the problem, of course, is to understand the problem itself.


Imagery: Unless otherwise noted, the images in this article are from the Polar Geospatial Center and the United States Geological Survey. The referenced study is “High-resolution elevation models of Larsen B glaciers extracted from 1960s imagery,” authored by Ryan North and Timothy T. Barrows.

Leave a Comment