Using best-track data from 1979-2015 to define the tracks of the storm centers, we identify storms that undergo ET by means of their paths in the cyclone phase space (CPS), calculated from geopotential height fields in reanalysis datasets. Two reanalyses are employed and compared for this purpose, the Japanese 55-year Reanalysis (JRA-55) and the ECMWF Interim Reanalysis (ERA-Interim). The results are used to study the seasonal and geographical distributions of storms undergoing ET, inter-basin differences in the statistics of ET occurrence, and the differences between the ETs defined by CPS and those defined by the 'extratropical' labels (determined subjectively by human forecasters using a wider range of data) in the best-track archives.
About 50% of all storms in the North Atlantic and the Western North Pacific undergo ET. In the southern hemisphere, ET fractions range from about 20% in the South Indian Ocean and the Australian region to 40% in the South Pacific.
The North Atlantic and Western North Pacific exhibit somewhat different seasonal cycles, with the probability of ET maximizing later in the North Atlantic, but having a local minimum in the earlier part of the peak season in both basins. Southern hemispheric basins have much less pronounced seasonal cycles.
The classification of ET storms based on JRA-55 agrees better with the best-track data than the ERA-Interim classification. In the North Atlantic and the Western North Pacific, the differences are small and both reanalyses achieve F1 performance scores of at least 0.8, but JRA-55 has a higher classification skill in all other basins.
Due to the global scope and consistent methodology, the results presented are well suited to serve as a benchmark for other studies including research on ET under climate change scenarios.