Mapping Built-up Areas in the Arctic at 10 m by Fusing Sentinel-2 and VIIRS Nighttime Lights Data

In recent years, climate change and globalization have spurred urbanization and infrastructure development in the Arctic. Yet climate change, especially the associated degradation of permafrost, also poses serious threats to the built environment and the sustainable development of Arctic cities and communities. Mapping built-up areas in the region using remotely sensed imagery is therefore crucial for monitoring urban and infrastructural expansion and for understanding the impacts of rapid permafrost thaw on urban environments at high northern latitudes. Although there are several global high-resolution built-up area datasets, the thematic accuracy of these datasets in Arctic is limited. The low spatial resolution of some of these datasets also results in an inability to capture certain small cities in the Arctic. To resolve these gaps, we develop a built-up area extraction method by combining Sentinel-2 and VIIRS nighttime lights data to map Arctic built-up areas from 2016-2018 at a resolution of 10 m. This method allows us to reveal the spatial patterns of Arctic urbanization in several representative areas (e.g. in Russia and Nuuk in Greenland) at an unprecedented spatial resolution. The built-up area mapping method we develop exhibits 1) high efficiency thanks to its selection of automatic training samples and determination of adaptive thresholds, and 2) high accuracy as a result of its elimination of interference from non-built-up areas, particularly bare land. By mapping the dynamics of built-up areas together with the distribution of Arctic permafrost, we are able to highlight the potential risks that rapid permafrost degradation presents to Arctic cities and infrastructure. Ultimately, this analysis will deepen understanding of Arctic urban sustainability within a rapidly changing climate.