Abstract: In the Arctic (60°N–90°N), water vapor transport directly and indirectly affects aspects of the climate system, including the radiation balance and the dynamics of the hydrological cycle. Understanding the distribution and variations of water vapor transport is important for the study of Arctic climate change. Using ERA5, JRA-55, and MERRA-2 reanalysis data from 1980 to 2018, we analyzed the horizontal and vertical distributions and the monthly and seasonal variations of water vapor flux at 70°N. Our results show that the direction of water vapor transport at 70°N is affected by the distribution of land and sea. Water vapor transport is northward in the eastern Arctic, and alternates between southward and northward in the western Arctic. Water vapor transport in the eastern Arctic is higher than that in the western Arctic. The northward water vapor transport is the highest near the 900-hPa level. At 70°N, the average northward water vapor transport is higher than the average southward water vapor transport. As a result, water vapor accumulates in the Arctic region. The northward water vapor transport is more active in the summer and autumn than in the winter and spring. The water vapor flux increases rapidly near the northward water vapor transport channel and in the North Atlantic region in the summer and autumn. Variations in the temporal and horizontal distributions of water vapor intrusion at 70°N indicate that intrusion mainly occurs near the northward water vapor transport channel, and is more strongly affected by specific humidity than by meridional wind. As a result, there were large fluctuations in the annual average number of water vapor intrusions after 2008. We computed the correlation coefficients between the IGRA2 observation dataset and each of the three reanalysis datasets. The correlation coefficient between IGRA2 and MERRA-2 is positive and is larger than the coefficient between IGRA2 and ERA5 and the coefficient between IGRA2 and JRA-55. Accuracy of MERRA-2 is also higher than that of ERA5 and JRA-55. We conclude that MERRA-2 is the most suitable dataset for the study of water vapor transport in the Arctic.

Key words: Arctic, reanalysis, water vapor transport, temporal and spatial distributions