Spatiotemporal variations of extreme events in surface mass balance over Greenland during 1958–2019.

Greenland surface mass balance (SMB) is undergoing dramatic change due to the amplified Arctic warming, with more frequent record‐breaking melt events. To comprehensively understand the behaviour of the Greenland ice sheet, we develop a suite of indices to examine the spatiotemporal variations of extreme events in SMB over Greenland during 1958–2019 based on the RACMO2.3p2 model outputs. We illustrate that the climatological distributions of extreme ablation‐dominated events (i.e., extremes with SMB < 0) share large similarity with the mean SMB in terms of intensity and frequency, showing a coastal‐to‐inland decreasing pattern. This pattern holds for the intensity of the extreme accumulation‐dominated events (i.e., extremes with SMB > 0), but not for the frequency which increases from the coastal to inland Greenland. Regarding the temporal evolution, the intensity and frequency of extreme ablation‐dominated events show a decreasing trend over Greenland during 1958–1978, whereas the trend increases afterwards, especially during the last two decades (2001–2019). However, these trends fluctuate regionally across the GrIS both in terms of magnitude and sign, with the most pronounced variations occurring in southwest Greenland. In contrast, extreme accumulation‐dominated events show, on average, a long‐term increasing trend from 1958 to 2019 in terms of intensity and frequency. However, obvious spatial fluctuations exist across the GrIS during 1958–2019, especially between the southwest and northeast Greenland where an opposite trend is observed. Additionally, the variations of SMB extremes in boreal summer are linked with the changes in regional temperature and precipitation and the associated atmospheric circulations, and the dominating factor varies with different extreme indices and time intervals. Our results may advance our understanding on SMB variability over Greenland and contribute to the design of future Greenland observational networks. [ABSTRACT FROM AUTHOR]