IceBoost v2.0: a machine learning model quantifies the volume of all the world's glaciers

Based on more than 7 million ice thickness measurements, the model estimates the volume of ice outside the ice caps at 150,000 km³, or 32.3 cm of potential sea level rise.

A team led by the Ca' Foscari University of Venice, with the Institute of Polar Sciences of the Italian CNR, published in Scientific Data A machine learning model called IceBoost v2.0, designed to reconstruct the thickness and ice volume of the planet's glaciers. The work is conducted by physicist Niccolò Maffezzoli.

The model was driven over 7 million measurements of thickness on glaciers around the world, crossed with 26 physical and geometric variables — slope and curvature of the terrain, ice flow rate, temperature. It estimates that glaciers outside Greenland and Antarctica contain about 150,000 cubic kilometres of ice, equivalent to 32.3 centimetres of average sea level rise.

The contribution is mainly due to the distribution: glacier ice distribution by glacier is up to 40% more consistent with field observations than previous estimates. The work also highlights areas where measures are sorely lacking — Himalayas, Karakoram and Patagonia. The study also includes teams from the University of California in Irvine, the Jet Propulsion Laboratory at NASA, Dartmouth College and the University of Copenhagen.