Using 20th Century reanalysis data to examine Northern Hemisphere storm track trend in the 20th Century

Several studies, based mainly on NCEP-NCAR and ERA40 data, suggested that Northern Hemisphere storm track activity has increased during the second half of the 20th Century. However, other studies have pointed out that NCEP-NCAR and ERA40 reanalysis data may contain spurious trends or jumps due to changes in the observing system. Since the 20th Century Reanalysis only assimilates surface pressure observations, it is expected to be less affected by changes in the observing system, especially over the second half of the 20th Century. In this study, 300 hPa meridional velocity variance, computed based on 20th Century reanalysis data, are compared to those computed based on rawinsonde observations, to assess the quality of the 20th Century reanalysis. Note that rawinsonde wind observations are not assimilated into 20th Century reanalysis data and thus can serve as an independent data set to assess its quality. Our results show that the 1958-1999 trend in band-pass filtered meridional velocity variance derived from 20th Century Reanalysis data is more consistent with that derived from rawinsonde observations than trends derived based on NCEP-NCAR and ERA40 data, despite the fact that rawinsonde wind observations are assimilated in the NCEP-NCAR and ERA40 data. Our results suggest that the increasing storm track trend found by previous studies in the winter North Pacific is probably spurious, while the increasing trend in the Atlantic is statistically significant but is most likely less that those estimated based on NCEP-NCAR and ERA40 reanalysis data. Currently, we are examining trends in surface cyclone statistics derived based on 20th Century reanalysis data and these statistics will be compared to trends derived based on other reanalysis data.