Consequently, Landy and his team drew on artificial intelligence to analyse past satellite data and determine when the satellites detect ice and when they detect ocean. By combining a deep learning method and numerical simulations, for the first time they succeeded in using the data to measure sea-ice thickness from space, even during the summer melting phase, with sufficient accuracy. The result: the first sea-ice thickness dataset that covers not only the entire Arctic, but also the entire year.
Comparing their findings with real-world observations allowed the group to gauge how accurate their new method was. In this regard, a great deal of the observational data they drew upon came from the IceBird campaigns conducted by the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI). “We use the AWI’s polar research aircraft to gather information on the composition and characteristics of the Arctic ice, and how they change over time,” says Dr Thomas Krumpen, a climate researcher at the AWI and co-author of the study. “The aerial data is highly precise and covers large expanses. That makes it particularly well suited for validation purposes.”
For a better understanding of processes at work in the Arctic, more precise forecasts and safer shipping
The new satellite data on summer sea-ice thickness is now to be integrated into sea-ice forecasts, so that accurate statements concerning the summer ice extent and volume can be made much sooner. “Now that we have year-round, Arctic-wide data, we’re gaining wholly new insights into interactions between the atmosphere and ocean,” says Krumpen. In addition, the outcomes are of considerable value for shipping in the Arctic, as well as future weather and climate forecasts: “Using the new satellite data, we are finally able to make sea ice forecasts informed by the ice thickness, not only for the winter, but also for the summer. This will reduce safety risks for ships and fishing,” says Landy. “We can also predict whether there will be ice or not at a given location in September, by measuring the ice thickness in May.” According to co-author Professor Michel Tsamados from University College London, the new data could not only improve short-term weather forecasts for the middle latitudes, but also long-term climate projections.