The volume of the planet's glaciers estimated by a new AI model

Scientists using artificial intelligence have calculated that Earth's glaciers contain approximately 150,000 cubic kilometers of ice, but they are losing mass at an unprecedented rate. The findings come from a new model developed by researchers at Ca' Foscari University in Venice and the National Research Council (Cnr) in Italy. This model, called IceBoost v2.0, combines millions of data points with physical and geographical variables to reconstruct glacier thickness and volume more accurately than previous methods. The study was published in Scientific Data, a journal under Nature. The research team trained the model using over seven million measurements, allowing them to estimate the volume and distribution of ice across all glaciers globally, excluding Antarctica and Greenland. A web application has been created to visualize these data, offering a detailed look at each glacier’s characteristics. According to the Cnr, the model integrates 26 different variables, including terrain slope, curvature, ice velocity, and temperature, which help build a point-by-point reconstruction of ice thickness and volume. The results show that glaciers worldwide hold roughly 150,000 cubic kilometers of ice. If this ice were to melt completely, it would raise global sea levels by approximately 32.3 centimeters. This figure aligns with earlier studies but offers a more accurate representation of individual glacier distributions, improving estimates by up to 40% compared to field measurements. One notable example is the Geikie Plateau in eastern Greenland, where the model suggests twice as much ice as previously recorded, reaching depths of up to two kilometers. The study highlights the rapid loss of ice mass from glaciers around the world. While the exact rate of melting varies by region, scientists warn that the pace is accelerating due to rising temperatures linked to climate change. The implications of such losses extend beyond sea-level rise, affecting freshwater availability, ecosystems, and human settlements dependent on glacial runoff. The model provides a critical tool for improving climate projections and water resource management strategies. Researchers emphasize that the accuracy of IceBoost v2.0 allows for better predictions of future changes in glacier systems. By integrating extensive datasets with advanced machine learning techniques, the model helps bridge gaps in understanding how glaciers respond to environmental pressures. This approach enables more precise modeling of potential scenarios, aiding policymakers and planners in preparing for the impacts of climate change. The collaboration between Ca' Foscari and Cnr underscores the growing role of interdisciplinary research in addressing complex environmental challenges. The development of tools like IceBoost v2.0 reflects advancements in both computational science and glaciology, offering new insights into one of the planet’s most vulnerable components. As the model continues to evolve, it will likely play a key role in shaping international efforts to monitor and mitigate the effects of global warming on polar and mountainous regions. The findings are expected to influence ongoing discussions within scientific communities and policy-making bodies regarding climate action. Future research will focus on refining the model further and expanding its applications to other cryospheric elements, such as ice sheets and permafrost. Scientists anticipate that the integration of additional data sources and improved algorithms will enhance the model’s predictive capabilities, providing even clearer guidance on the trajectory of global ice loss.