A Comparison Between Two Kinds Of Reanalysis Data In The Stratospheric Temperature

The ECMWF and NCEP/NCAR reanalysis datasets are compared of globalmonthly mean temperature field in lower stratosphere(150～10hPa)based on Lorenzcirculation resoluting. Both ECMWF and NCEP datasets are available on a 2.5°by2.5°grid with time series spanning 1958.1～2001.12.Fistly, in order to analyse the different of spatial distribution features betweenthe two reanalysis datasets on latitude-altitude section, the zonal and climatologicalnormals discrepance of these datas were did a Student's t test, and the zonal meantemperature's interannual variance discrepance of these datas were did a F test. Thenlower stratospheric stationary waves on southern and northern hemisphere in DJF orJJA were discussed respectively. What difference between the two reanalysis datasand which one more logical we needed to attention. Finally, by analyzing thecharacters of temperature anomaly on global,hemispheres and tropic,we may find outwhich reanalyses can respond enough to that the major eruption of threevolcanos(Agung in 1963, Elchichon in 1982 and Mt.Pinatubo in 1991) the latest 50years arose out of elevated temperature in lower strarosphere.The results are given as follow: (1)in DJF there is a cold center on arctic, whereclimatological normals of ECMWF reanalyses are lower than ones of NCEPreanalyses and there is a warm center on Antarctica, where climatological normals ofECMWF reanalyses are higher than ones of NCEP reanalyses.In JJA, contrarily.Whether in DJF or in JJA,the zonal available temperature of 30°S～30°N is a coldcenter in 100hPa.In this area, the climatological normals of ECMWF reanalysis arelower than those of NCEP reanalysis significantly and the interannual variances ofECMWF reanalysis are smaller than the ones of NCEP.(2) Below the level of 30hPa, the stationary wave of ECMWF reanalyses is similar to the one of NCEPreanalyses. But from 30hPa to the top of lower stratosphere, the stationary wave ofECMWF reanalyses is comparatively rational. Because in NCEP stationary wavethere were several small-scale disturbances what is data noise produced in datahandling. (3)Two kinds of reanalyses datasets all reveal that 50～70hPa wereinfluenced by volcanic activity most markedly. But the temperature anomalies of thetwo datasets are different on the level of 100hPa and 10hPa. Talking one with another,the effect of Pinatubo volcano were the markest on global and hemispheres and theone of Agung volcano were the markest on tropic.