The NCEP/NCAR Reanalysis dataset provides a valuable resource for examining the past 50+ years (1948-) atmospheric patterns. We create separate composite anomaly fields for El Nino and La Nina periods using the NCEP/NCAR Reanalysis data and compare the wind and sea level pressure anomaly patterns found with previous composite anomalies created with the COADS data set (Larkin and Harrison, 2002, 2001; Harrison and Larkin, 1996, 1998). We examine composite anomalies for the period Yr(-1) through Yr(+1) over the globe, and find both substantial areas of agreement as well as regions with large and significant departures.

For both El Nino and La Nina events, the best agreements are found in the Tropical Pacific during the peak of the event lifecycle, with significant differences occuring outside of the Pacific waveguide in all variables during all phases of the event. Meridional wind shows the least agreement, with zonal wind showing the best agreement. Differences between the data sets include both the additional presence of anomaly patterns in the NCEP/NCAR data that are not found in the COADS composites, as well as the absence of anomaly patterns determined to be statistically significant in the COADS composites. Focusing on anomaly patterns determined to be "robust" in the COADS data in that they are substantially present in all or nearly all the El Nino / La Nina events, we find that a large number (1/3 +) of these robust anomalies show significant differences between the two data sets, even in the peak of the event in the tropical Pacific. These differences include such dynamically suggestive anomalies as the lack of previously seen northerly anomalies in the tropical Pacific ITCZ region during El Nino events. In general, the NCEP/NCAR Reanalysis composites show a greater spatial detail and more tendency to have wavetrain like anomalies than were found in the COADS composites.