Extreme Precipitation in the Eastern Canadian Arctic and Greenland: An Evaluation of Atmospheric Reanalyses

• Published in: Frontiers in Environmental Science
• Main Authors: Nicole A. Loeb, Alex Crawford, Julienne C. Stroeve, John Hanesiak
• Format: Article in Journal/Newspaper
• Language: English
• Published: Frontiers Media S.A. 2022
• Subjects: extreme precipitation; atmospheric reanalysis; Arctic; climate; Greenland; Canadian Arctic; Environmental sciences; GE1-350; Merra
• Online Access: https://doi.org/10.3389/fenvs.2022.866929; https://doaj.org/article/f3c77614667345c3b37e4f45ea023221

Extreme precipitation events are becoming more common in the Arctic as the climate warms, but characterizing these events is notoriously challenging. Atmospheric reanalyses have become popular tools for climate studies in data-sparse regions such as the Arctic. While modern reanalyses have been shown to perform reasonably well at reproducing Arctic climate, their ability to represent extreme precipitation events has not been investigated in depth. In this study, three of the most recent reanalyses, ERA-5, MERRA-2, and CFSR, are compared to surface precipitation observations in the Eastern Canadian Arctic and Greenland from 1980 to 2016 to assess how well they represent the most intense observed events. Overall, the reanalyses struggled to match observed accumulations from individual events (−0.11 ≤ r ≤ 0.47) but matched the observed seasonality of precipitation extremes. The region with the strongest match between observations and reanalyses was Southwest Greenland. Performance varies by event, and the best match between reanalyses and station observations may have a spatial/temporal offset (up to one grid cell or 1 day). The three products saw similar performance in general; however, ERA-5 tends to see slightly higher correlations and lower biases than MERRA-2 or CFSR. Considering the limitations of in situ observations, these results suggest that the reanalyses are capable of representing aggregate extreme precipitation (e.g., seasonal or annual time scales), but struggle to consistently match the timing and location of specific observed events.