Extreme temperature and precipitation events in Alaska are examined in an ensemble of global climate models (GCMs) and an atmospheric reanalysis. Extreme monthly maximum and minimum temperature and the monthly maximum 5-day precipitation amount are evaluated for a 30-year historical period and two 30-year future time slices (2050s and 2080s). Although biases exist, models capture the spatial pattern and seasonality of the extremes depicted by the ERA-40 reanalysis. Discrepancies between station data (Anchorage, Fairbanks and Barrow) and GCMs/reanalysis are larger than the model-reanalysis differences, and are consistent with (1) surface elevation differences arising from the models’ resolutionand (2) gauge undercatch of precipitation in the station data. GCMs project future changes that are 2-4 times larger than the across-model standard deviations. The largest changes projected by the GCMs are significantly different from the historical mean at the 95% confidence level. Changes in extreme minimum temperature and extreme 5-day precipitation projections are larger than changes in means. The extreme minimum temperatures are projected to increase 2-3 times as much as the extreme maximum temperatures in all seasons except summer, with the largest increases of extreme minima in coastal regions. By the 2080s, the increases in all three extremes indices are twice as large in the Representative Concentration Pathway (RCP) 8.5 as in the RCP 4.5 scenario. The magnitude of the projected increase of maximum 5-day precipitation is largest in southern and inland areas, although the percentage increase is largest in the north. In the RCP 8.5 simulations, the inter-annual variability of extreme temperatures narrows by the end of the century, most notably in autumn. Record-breaking 5-day precipitation events become more common in the RCP 8.5 than in the RCP 4.5 scenario.
Bennett, K. E. and J. E. Walsh. 2014. Spatial and temporal changes in indices of extreme precipitation and temperature for Alaska. International Journal of Climatology. 35: 1434-1452. http://onlinelibrary.wiley.com/doi/10.1002/joc.4067/epdf.