Arctic infrastructure was built on frozen ground that’s starting to thaw – drones can help spot the permafrost changes coming

The town of Wainwright in Alaska appears stable from above, but beneath its surface lies a hidden threat: large bodies of ice within the frozen ground, known as permafrost. Permafrost is ground that remains at or below 0°C (32°F) for at least two consecutive years. As Arctic temperatures rise two to three times faster than the global average, this ice is melting, causing the ground to sink and destabilize. Roads, buildings, and pipelines built on this thawing ground can tilt, crack, or collapse. In Alaska alone, permafrost thaw could cost between $37 billion and $51 billion in building and road damage by 2025. Along the coast, melting ice exposes the ground to erosion, bringing the ocean closer to homes and infrastructure. For example, in Utqiaġvik, Alaska, a bluff partially collapsed in 2023 and 2024, narrowing the distance between eroding land and buildings.

Permafrost thaw doesn’t affect all areas equally. Some locations may contain more ice than others, and the presence of buildings or roads can alter how snow and water move through the ground. When ice melts, the ground loses its ability to support weight, leading to uneven sinking. This process is called subsidence. Along coastlines, thawing permafrost makes shorelines more vulnerable to waves and erosion. For instance, in Utqiaġvik, waves eroded the bottom of a bluff while thaw weakened the ground above, causing the edge to slump. The uneven nature of permafrost thaw means that areas just a city block apart can have vastly different risks. This unpredictability makes it challenging to plan for long-term infrastructure stability.

Traditional methods to study permafrost, like drilling boreholes, are expensive, labor-intensive, and disruptive. These methods involve taking samples of the ground below, which is difficult in frozen soil. Instead, scientists use geophysical imaging techniques to map the ground without direct contact. One such method is very low frequency electromagnetic surveying (VLF-EM), which measures how the ground responds to electromagnetic signals from distant transmitters. A lightweight sensor carried by a drone can survey wetlands and tundra that are hard to reach. VLF-EM is especially useful along coasts because salty, unfrozen water conducts electricity better than ice-rich frozen ground, helping identify thawed areas. While this method doesn’t provide a perfect underground picture, it can highlight where conditions have changed, guiding engineers to areas needing closer inspection.

Infrastructure in the Arctic often serves for decades, so knowing current permafrost conditions isn’t enough. Engineers need to forecast how thaw will affect ground stability over time. Researchers have developed models to project two key factors: how much the ground will settle and how much load a foundation can support. A 2026 study examined permafrost stability across Alaska’s Arctic Coastal Plain, which includes Prudhoe Bay and much of the state’s oil industry. If greenhouse gas emissions remain high, the study found that less than 10% of infrastructure is at risk until mid-century. However, between the 2060s and 2080s, the risk rises rapidly. By the 2090s, about 80% of buildings, 60% of roads, and 90% of pipelines are projected to be at risk of damage from sinking land.

To better understand and predict permafrost changes, researchers are using digital twins—computer models that replicate real-world environments and aging infrastructure. These models combine heat-transfer physics and machine learning to adjust predictions as new data arrives. For example, a team used fiber-optic cables to collect temperature and seismic data from a 330-foot (100-meter) section of a road embankment in Utqiaġvik. The digital twin updated its predictions based on this data, improving forecasts of permafrost temperature changes and infrastructure performance. The goal is to use drones for mapping, sensors for tracking changes, and digital twins for forecasting how these changes will impact infrastructure over time. This approach provides communities with tools to monitor hazards, plan repairs, and adapt to future conditions.