| Summary: | The precipitation over the Arctic region is a difficult quantity to determine with high accuracy, as the in situ observation network is sparse, and current climate models, atmospheric reanalyses, and direct satellite-based precipitation observations suffer from diverse difficulties that hinder the correct assessment of precipitation. We undertake a proof-of-concept investigation into how accurately optical satellite observations, namely Sentinel-2 surface-reflectance-based grain-size-connected specific surface area of snow (SSA), and microwave-based snow water equivalent (SWE) estimates can detect snowfall over the Arctic. In addition to the satellite data, we also include ERA5-Land SWE data to support the analysis. Here, we chose a limited area (a circle of 100 km radius around Luosto radar located in Northern Finland) and a short time period (covering March 2018) to test these data sources and their usability in this precipitation assessment problem. We classified differences between observations independently for SSA and SWE and compared the results to the radar-based snowfall information. These initial results are promising. Situations with snowfall are classified with high recalls, 64 % for the satellite-based SWE, 77 % for ERA5-Land-based SWE, and around 90 % for SSA compared to radar-based data. Cases without snowfall are more difficult to classify correctly using satellite-based data. The recall values are 34 % for satellite-based SWE and vary from almost 60 % to over 70 % for SSA. SWE from ERA5-Land has the highest recall value for cases without snowfall, 80 %. These results indicate that optical and microwave-based satellite observations can be used to detect snowfall events over the Arctic.
|
|---|