An automated algorithm to detect Cold Air Outbreaks in ERA-40 and WACCM

Numerous Cold Air Outbreaks (CAOs) occur over North America, Europe and Asia each winter. Extreme CAOs result in the loss of life and extensive damage to crops. Thus, it is of interest to estimate changes in CAO characteristics as a result of climate change.To identify CAOs and compile statistics, an automated algorithm has been developed using ECMWF Reanalysis (ERA-40) data. The algorithm was applied to 45 years of ERA-40 temperature data from September 1st,1957 to August 31st, 2002, focusing on days between October 1st and May 1st. Gridpoints where the temperature is 1.5 standard deviations below the 31 day climatological mean were flagged as being part of a CAO event. The fidelity of the algorithm was checked via visual inspection of the data; the algorithm consistently identified CAOs correctly over the entire data record. The algorithm was then applied to 30 years of Whole Atmosphere Community Climate Model (WACCM) version 3 output for comparison to ERA-40 results. The CAO climatologies are used to look for changes in occurrence frequency, strength, geographical extent, and dependence on hemispheric oscillations such as the North Atlantic Oscillation.