Near-surface air temperature extremes in the Forlandsundet region (northwest Spitsbergen, 2010–2015)

The Arctic has undergone rapid climatic change over recent decades; however, the spatial variability of temperature extremes in complex coastal, tundra and glacial environments remains insufficiently quantified. This study investigates near-surface air temperature extremes in the Forlandsundet region (northwest Spitsbergen) over the period 2010–2015, based on year-round observations from six measurement sites representing distinct topoclimatic settings, including a mountain ridge, a glacier and its marginal zone, a moraine, a tundra, and a coastal zone.

Temperature extremes were identified using a station-specific, monthly percentile approach, allowing for the characterization of extremes relative to local climatic conditions and seasonal variability. Warm extremes were defined by the 90th, 95th, and 99th percentiles, while cold extremes were determined using the 10th, 5th, and 1st percentiles of mean daily air temperature. This method enabled a consistent assessment of both the frequency and intensity of extreme thermal events across contrasting environments.

The analysis focuses on spatial differentiation of extremes and on the role of topography, surface characteristics, and elevation in shaping local thermal regimes. Additionally, the influence of large-scale atmospheric circulation was examined using the Niedźwiedź classification for Svalbard, which enabled the identification of synoptic conditions associated with extreme temperature events. The role of local meteorological factors, including cloudiness and wind conditions, was also investigated.

The results provide a comprehensive assessment of temperature extremes across diverse Arctic topoclimates and highlight the combined influence of synoptic forcing and local environmental controls. These findings contribute to a better understanding of fine-scale climate variability in polar regions and offer valuable insights for improving Arctic regional climate modelling.