Warm spells in the central Arctic during winter and spring are typically associated with intrusions of warm and moist air from the mid-latitudes (warm air intrusions hereafter). Two such events took place in April 2020. They passed the German icebreaker Polarstern, which was drifting in sea ice as part of the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) campaign. In both cases, the surface temperature measured on Polarstern rose from the normal April temperature of around –20°C to around 0°C. ECMWF forecasts supported targeted observations made during this period, which will in turn help the Centre to improve its ability to predict warm air intrusions.
MOSAiC is a one-year observational campaign linked to the World Meteorological Organization’s Year of Polar Prediction (YOPP) project. YOPP aims to understand atmospheric, ocean and sea-ice processes and advance prediction capabilities in polar regions. The MOSAiC campaign started in October 2019. YOPP and MOSAiC are led by the Alfred Wegener Institute (AWI, Germany), and the US National Oceanic and Atmospheric Administration (NOAA) coordinates forecast evaluation at MOSAiC.
Targeted observations
The third period of enhanced Arctic observations during YOPP took place in spring 2020 to complement the MOSAiC campaign. It took the form of a targeted observing period (TOP), which was different from the earlier special observing periods (SOPs) in 2018. In the TOPs, extra observations were only requested during selected meteorological situations of relevance for the Arctic. During this TOP, additional radiosondes were launched from different stations situated along warm air intrusions in order to shed light on the processes governing these situations. During the period of the warm air events, when Polarstern was located north of Svalbard, four radiosondes a day were launched at several upstream locations and seven a day on board Polarstern.
Although the two warm air intrusion events on 16 and 19 April were close in time, they were associated with different synoptic patterns. In the first event, the warm air was pushed to the northeast in front of a trough over Scandinavia, while in the second event the warm air was transported over the Atlantic on the western side of a ridge that developed over Scandinavia. The second case is a more typical flow configuration for creating warm conditions in the Atlantic sector of the Arctic.
Valuable guidance
Various ECMWF forecast products were used in the planning for the TOP, such as the Extreme Forecast Index (EFI) for temperature and water-vapour flux, upper-level flow forecasts and meteograms for the Polarstern location. Evaluating the performance of all forecasts for 2‑metre maximum temperature valid on 19 April, the ensemble had a clear signal 7–8 days in advance about warmer than normal conditions at the location of Polarstern. The EFI product for 2‑metre maximum temperature from five days before 19 April also flagged up the risk of unusually high temperatures. By 19 April, the first warm air intrusion was being advected to the east (north of Russia) and the second one was north of Svalbard, as predicted. The second event was associated with large uncertainties due to the narrow nature of the warm-air feature. For example, in the forecast from 15 April, while most of the members indicated a very high daily maximum temperature, many ensemble members ended the warm spell too early at the Polarstern location, resulting in massive temperature errors during most parts of the day. Although there were large uncertainties in the timing and details of the event, the ECMWF forecasts gave valuable guidance for observation campaign planning.
Outlook
The data collected during the MOSAiC campaign and the YOPP special observing periods will be used for model evaluation and development at ECMWF. To correctly forecast warm air intrusions and associated impacts, several aspects such as the origin and nature of the air mass, synoptic conditions, interactions with sea ice and mixed-phase cloud processes need to be correctly captured. It is therefore necessary to understand the physical mechanisms governing such events in order to improve the forecasting system. To this end, observations such as the ones from MOSAiC and YOPP are essential.