May 12, 2021

Glacier melting: untangling local from global

An international team led by the LEGOS space geophysics and oceanography research laboratory has measured the evolution of all the world’s glaciers over 20 years. On average, they are losing 267 billion tonnes in ice mass a year, a cumulative loss of 4% in just 20 years. The authors confirm this phenomenon is accelerating globally as temperatures rise and explains certain slowdowns observed locally.

Upsala Glacier, Argentina. Credit: Étienne Berthier

From their 713-kilometre perch in Earth orbit, the cameras of Japan’s ASTER instrument (Advanced Spaceborne Thermal Emission and Reflection Radiometer) on NASA’s Terra satellite have been capturing frequent images all over the surface of the globe for 20 years. These vast volumes of imagery have enabled an international team led by Romain Hugonnet and Étienne Berthier, glaciologists at the LEGOS space geophysics and oceanography research laboratory, to precisely measure glacier evolution. Each of the 220,000 catalogued glaciers on Earth—outside the polar ice caps—has its own specific features, but together they have been losing on average 267 gigatonnes every year for the last 20 years. This is the first time a team has been able to obtain a synoptic picture of changes in glacier mass over such a long period.

500 000 images to process, anomalies to explain

Images acquired by ASTER between 2003 and 2018 showing that the 800-km-long Upsala Glacier has retreated by 4 km.

redits: ASTER-NASA-JAXA-Étienne Berthier

In a warming world, how can the glaciers of the Karakoram mountain range in the Western Himalayas have gained mass up to 2015? Why are the glaciers of the Northern Atlantic still today melting faster than expected? The paper published this April in the review Nature gives a coherent explanation for temporally and spatially limited anomalies of this kind. The paper’s authors considered the two main factors likely to impact glaciers: temperature and precipitation. In a drying region, glaciers are no longer replenished by fresh snow accumulating at the surface. The lack of precipitation therefore causes them to lose mass more quickly. For example, the great drought of 2010 considerably speeded up the rate of glacier thinning in the Central Andes. In contrast, a local increase in precipitation explains the slowing of glacier mass loss observed over the last 10 years around the Northern Atlantic.

With an average of 40 observations from each of the 220,000 glaciers, the research scientists at LEGOS showed that at global scale precipitation is making a modest contribution to glacier evolution compared to temperature. Rising temperatures fuelled by global warming are accelerating glacier melt. The dataset spanning two decades revealed a mean annual loss of 227 gigatonnes between 2000 and 2005, rising to 298 gigatonnes over the 2015-2020 period.

It took eight years of research and the efforts of three doctoral researchers—Fanny Brun, Inés Dussaillant and Romain Hugonnet—working under the direction of Étienne Berthier to obtain these results. To begin with, the principle was fairly simple: “the idea was to map changes in elevation as we did in the 1950s using aerial photographs to compare glaciers,” says Berthier. Glacier thinning is then interpreted as a loss of mass. While the principle was simple, statistical processing was not nearly as straightforward. The scientists were no longer working with aerial photos, but rather 500,000 satellite images with a pixel resolution of 15 metres and numerous erroneous values to be filtered out.


Kongsfjorden glaciers in Svalbard, Norway, viewed by ASTER.
Credits: NASA-JAXA-Étienne Berthier

Glacier melting could have multiple consequences. The main risk relates to sea level rise, 20% of which can be attributed to glacier melting. Models for predicting coastal erosion, shoreline retreat and marine submersion due to sea level rise will need to be revisited in the light of accelerating glacier melt. Another major consequence is that when a glacier thins we lose a precious reserve of freshwater. In winter, the vast slab of ice stores water and releases it in summer, regulating the resource like a water tower.

When 267 gigatonnes of ice disappears, it’s the equivalent of 10 years of freshwater consumption in France—for electric power stations, agriculture and drinking water—that’s lost. In dry regions, these water reserves are crucial to populations. To improve governance of lands and better anticipate climate impacts, simulations incorporating these data would be more representative. Indeed, American, Austrian and Swiss research scientists have already fed them into their models.

For Étienne Berthier, the impact of these data in informing policy decisions amply justifies envisioning their long-term continuity. But how will we acquire new data when ASTER is about to cease operating after two decades in service? “No current or planned missions will fill this gap, so with CNES we’ve started dreaming of a follow-on,” he says—a dream in the form of a satellite able to observe Earth continuously with high-resolution stereo imagery, frequent revisits and open data.


  • Étienne Berthier

Research Director at LEGOS, Satellite Cryosphere team (CNRS/CNES/IRD/ Paul Sabatier University - Toulouse III)
Adress : Observatoire Midi-Pyrénées, 14 Avenue Edouard Belin, 31400 Toulouse
E-mail :
Tel. : + 33 (0)5 61 33 47 15

  • Anne Lifermann

Head of Cryosphere and Coasts, CNES
Adress : Centre National d'Etudes Spatiales 18 avenue Edouard Belin 31401 Toulouse Cedex 09, France
E-mail : anne.lifermann at
Tel. : +33 (0)5 61 28 21 43