New AI model reveals the volume of the world's glaciers
How much ice is stored in the world's glaciers? And where exactly is it located? A new study led by Ca' Foscari University of Venice, in collaboration with the Institute of Polar Sciences of the National Research Council of Italy (CNR-ISP), provides an updated global map of glacier ice volume.
Published in Scientific Data, the study introduces IceBoost v2.0, a machine-learning model developed by Niccolò Maffezzoli, physicist and researcher at Ca' Foscari University of Venice and affiliated with CNR-ISP. IceBoost v2.0 was trained on more than 7 million ice-thickness measurements collected from glaciers across the globe.
The model combines these observations with 26 physical and geometrical variables (including topographic slope and curvature, ice velocity, and temperature) to reconstruct, on a point-by-point basis, the ice thickness and thus the volume of every glacier included in the Randolph Glacier Inventory (RGI), the world's most up-to-date global glacier inventory, excluding the polar ice sheets.
An interactive web app allows users to explore the data from Maffezzoli's model for every glacier worldwide: https://nmaffe.github.io/iceboost_webapp/.
According to the new estimates, the world's glaciers contain approximately 150,000 cubic kilometers of ice, equivalent to 32.3 centimeters of global mean sea-level rise if they were to melt completely (excluding Antarctica and Greenland). While this figure is consistent with previous global estimates, IceBoost v2.0 provides a far more realistic estimate of each glacier's distribution, matching field observations up to 40% more accurately.
-
Svalbard archipelago. Credit: Ca' Foscari University of Venice -
Severnaya Zemlya (Russian Arctic). Credit: Ca' Foscari University of Venice -
Prince of Wales icefield (Arctic Canada North). Credit: Ca' Foscari University of Venice
One of the most striking examples is the Geikie Plateau in eastern Greenland, where the glacier is up to 2 kilometers (1.2 miles) thick. In this area, IceBoost v2.0 estimates nearly twice as much ice as previously reported.
"The distribution of glacier ice thickness is a fundamental variable for glaciological and climate models," Maffezzoli explains. "To predict how glaciers will evolve by 2100 and quantify their contribution to sea-level rise, we first need the most detailed possible picture of their current state. Researchers involved in the Glacier Model Intercomparison Project (GlacierMIP4), who are producing the next generation of glacier simulations to inform the IPCC's assessments of glacier evolution through 2100, will use IceBoost v2.0 as their sole representation of the present-day situation."
The new dataset also provides valuable guidance for future field campaigns and scientific investments. IceBoost v2.0 maps identify regions where model estimates are most reliable, as well as areas where additional observations are most urgently needed, including the Himalaya and Karakoram ranges and the major Patagonian ice fields.
The research also has significant implications for freshwater management. Glaciers sustain rivers, ecosystems, agriculture and local communities, supporting the livelihoods of around 1.9 billion people worldwide. Better estimates of glacier thickness and ice volume make it possible to produce more reliable projections of future freshwater availability, particularly in increasingly arid and desertifying regions such as parts of South America.
-
Geikie Plateau (coastal East Greenland). Credit: Ca' Foscari University of Venice -
Fedchenko glacier and Yazgulem Range (Pamir, Tajikistan). Credit: Ca' Foscari University of Venice -
Devon ice cap (Arctic Canada North). Credit: Ca' Foscari University of Venice
"Glaciers are highly complex systems. Although the governing physics is well understood, many of the parameters involved in the equations remain unknown, poorly constrained or extremely difficult to measure," Maffezzoli concludes. "Machine-learning models offer an alternative approach: They learn directly from data, generating predictions without imposing a predefined physical description.
"When sufficient data are available, this approach often proves highly effective. The future lies in hybrid models, where physical modeling and learning from experimental measurements work together to produce even more accurate estimates. We must move quickly: Glaciers at midlatitudes, including those in the Alps, are expected to disappear within the next few decades."
Publication details
Niccolò Maffezzoli et al, Machine-learned global glacier ice volumes, Scientific Data (2026). DOI: 10.1038/s41597-026-07744-9
Who's behind this story?
Swati Mestri holds a bachelor's degree in Electronics Engineering and has worked as a content editor since 2019. She has experience editing research documents across technology, health care, and materials science, and has a particular interest in technology and space. Full profile →
Bachelor's in mathematical biology, Master's in creative writing. Well-traveled with unique perspectives on science and language. Full profile →
Citation: New AI model reveals the volume of the world's glaciers (2026, August 7) retrieved 8 August 2026 from https://phys.org/news/2026-08-ai-reveals-volume-world-glaciers.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.