Comparison of Atmospheric Large-scale Patterns during two Warming Periods in Greenland in the last 100 years

The air temperature (AT) increased during the Early 20th Century Warming (ETCW), especially in the Arctic, with a similar trend as during the present warming period. This AT increase is observed while investigating the annual AT anomaly of historic observations provided by the Danish Meteorological Institute (DMI) and of the zonal average of Greenland based on reanalysis data (NOAA 20CRv3). 

We define two distinct warming periods (1922–1932 and 1993–2007) for Greenland with a continuous increase in the AT anomaly. The increase is the largest at the northernmost observations in Upernavik and the smallest at the easternmost observations in Tasiilaq. The zonal average trend (Sen's slope) of AT increase in Greenland is 0.1°C/year in both periods, exceeding the global AT trend. Examining the spatial distribution of the AT trend in the reanalysis data during the warming periods reveals a warming hotspot in the sea in front of the West Coast of Greenland, which is more dominant in the second period. Nonetheless, the positive trend is rather homogeneous over Greenland, indicative of large-scale influences rather than localized phenomena. This motivates our study to analyse and compare the structure of atmospheric large-scale patterns (LSP) during these two warming periods. 

To do this, we use an unsupervised self-organizing maps (SOM) algorithm to highlight prevalent LSPs based on the reanalysis of the geopotential height of 500hPa. SOM is an artificial neural network used for clustering data into distinct groups, so-called nodes, by reducing its dimensionality. In the first approach to compare both periods, the frequency of the nodes is evaluated, meaning comparing how often a specific prevalent LSP defined by SOM occurs in the one and the other warming periods. A preliminary result is that there are significant differences in the occurrence of the nodes. Further exploration of the difference in node frequency and setting them into a meteorological context are the primary objectives of this study. 

Additionally, we aim to establish links between LSP and anomalies of atmospheric variables (such as air temperature) to investigate whether similar LSP are accountable for similar deviations. This will deepen our understanding of the atmospheric dynamics during Greenland's warming periods, which affect the cryosphere.