A new study highlights the economic risks that melting permafrost poses to Arctic communities. Permafrost is ground that remains frozen for at least two years straight. It covers about a quarter of the Northern Hemisphere, including large parts of Alaska, Canada, and Siberia. As global temperatures rise, this frozen ground is melting, releasing stored greenhouse gases and damaging infrastructure. The study, published in Earth's Future, focused on the costs of rebuilding infrastructure affected by permafrost thaw, especially in Siberia, where many multistory apartment buildings were built during the Soviet era to support resource extraction. The research team, led by Elias Manos, a doctoral candidate in Natural Resources and the Environment, and Chandi Witharana, an assistant professor in the same department, used a three-dimensional approach to assess building damage. Previous studies had underestimated the costs by only considering buildings' two-dimensional structures. By mapping buildings in 3D, the team found that the projected cost of rebuilding due to permafrost melt by the middle of the 21st century could be as high as $261 billion—more than double the previous estimate of $110 billion. This significant increase is due to the much higher material costs of rebuilding multistory buildings compared to single-story ones. To achieve this detailed analysis, the researchers used a combination of satellite images and artificial intelligence. They trained a deep learning model on a large dataset of satellite images to identify buildings and classify them by use type, such as residential or commercial. This allowed them to estimate building heights. They also used the ArcticDEM dataset, which provided elevation data, to calculate the height of each building by comparing the elevation at its base and top. Finally, they applied building codes from the U.S., Canada, and Russia to estimate the average height of each floor in residential buildings. This method was not used for nonresidential buildings due to the wide variation in their structures. The study's findings are crucial for Arctic communities and governments developing adaptation plans. The researchers used a model of permafrost degradation developed by one of their collaborators to estimate future damages. However, the study did not include demographic data, which would help identify which buildings are likely to remain in use. This is an area for future research. The researchers have shared their findings through the Permafrost Discovery Gateway, an international platform. Their method can also be applied to industrial structures, offering valuable insights for industries operating in the Arctic, despite the region's low population.