In this work, some fundamental characteristics of the Southern Hemisphere (SH) storm tracks are investigated by using the reanalyses, observational data as well as IPCC 20th Century simulations. Large discrepancies exist between the NCEP/NCAR and ECMWF ERA40 reanalyses in depicting the strength of the SH storm tracks, one of its key and fundamental properties. By comparing the reanalyses with radiosonde and satellite retrieved temperature (SATEMP) observations, it is found that the SH storm tracks in the NCEP/NCAR reanalysis is biased weak because the NCEP/NCAR reanalysis assimilates SATEMP (instead of the raw radiance) whose variance is biased low. Furthermore, the first quantitative investigation on the real strength of the SH storm tracks is carried out by using a new independent dataset: GPS radio occultation data obtained by the Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) project. It is found that the strength of the SH storm tracks in the NCEP/NCAR reanalysis is biased low by about 25%, and that in the ERA40 is closer to that observed, but still biased low by about 5%-10%. The trend of the SH storm tracks is another important issue, and again the two reanalyses do not agree on it. The NCEP/NCAR reanalysis shows a constant increasing trend since 1979, while the ERA40 shows very weak increasing trend but mainly a poleward shift. Radiosonde observations at stations close to the SH storm tracks, and satellite derived precipitation from the Global Precipitation Climatology Project are examined, with the results showing that the trends in these observational datasets are more consistent with that in the ERA40. Furthermore, it is found that the trend of the SH storm tracks and that of the SH mean flow is more dynamically consistent in the ERA40 than that in the NCEP/NCAR reanalysis. Therefore, it is concluded that the long term change of the SH storm tracks presents as mainly a poleward shifting with some slight strengthening. The implication of the introduction of massive satellite observations since late 1970s to the long term trend of the SH storm tracks are finally discussed.