An analysis of winter rain-on-snow climatology in Svalbard

• Published in: Frontiers in Earth Science
• Main Authors: Hannah Vickers, Tuomo Saloranta, Morten Køltzow, Ward J. J. van Pelt, Eirik Malnes
• Format: Article in Journal/Newspaper
• Language: English
• Published: Frontiers Media S.A. 2024
• Subjects: rain-on-snow; remote sensing; polar climate; climate models; snow cover; Science; Q; Svalbard; Spitsbergen
• Online Access: https://doi.org/10.3389/feart.2024.1342731; https://doaj.org/article/40ea452c3d5646efa839bfa760ad57c1

Rain-on-snow (ROS) events are becoming an increasingly common feature of the wintertime climate Svalbard in the High Arctic due to a warming climate. Changes in the frequency, intensity, and spatial distribution of wintertime ROS events in Svalbard are important to understand and quantify due their wide-ranging impacts on the physical environment as well as on human activity. Due to the sparse nature of ground observations across Svalbard, tools for mapping and long-term monitoring of ROS events over large spatial areas are reliant on remote sensing, snow models and atmospheric reanalyses. However, different methods of identifying and measuring ROS events can often present different interpretations of ROS climatology. This study compares a recently published Synthetic Aperture Radar (SAR) based ROS dataset for Svalbard to ROS derived from two snow models and a reanalysis dataset for 2004–2020. Although the number of ROS events differs across the datasets, all datasets exhibit both similarities and differences in the geographical distribution of ROS across the largest island, Spitsbergen. Southern and western coastal areas experience ROS most frequently during the wintertime, with the early winter (November–December) experiencing overall most events compared to the spring (March–April). Moreover, we find that different temperature thresholds are required to obtain the best spatial agreement of ROS events in the model and reanalysis datasets with ground observations. The reanalysis dataset evaluated against ground observations was superior to the other datasets in terms of accuracy due to the assimilation of ground observations into the dataset. The SAR dataset consistently scored lowest in terms of its overall accuracy due to many more false detections, an issue which is most likely explained by the persistence of moisture in the snowpack following the end of a ROS event. Our study not only highlights some spatial differences in ROS frequency and trends but also how comparisons between different datasets can ...