In order to assess the agreement among three First GARP Global Experiment (FGGE) Level IIB analysis data sets produced by the European Centre for Medium Range Weather Forecasts, the Geophysical Fluid Dynamics Laboratory, and the Goddard Laboratory for Atmospheres, the intercomparisons of daily values of large-scale vorticity &zgr;, divergence Δ, temperature T, static stability &Ggr;, vertical motion ω, and diabatic heating rate Q/CP are performed during the period January 26 to February 11, 1979. The &zgr;, Δ, and ω fields are calculated from the velocity components u and v after interpolation of Level IIIb analyses onto the National Center for Atmospheric Research Community Climate model, version 0B (NCAR CCM0B) grid. The Q/CP is obtained as a residual of the thermodynamic energy balance. Using standard descriptive and comparative statistics, we intercompare the three analyses in three geographic regions: NH (60¿N--30¿N), TR (30¿N--30¿S) and SH (30¿S--60¿S).

The agreement among the &zgr; analyses is excellent, though a slight disagreement is found in the lower troposphere of the TR and SH. However, the agreement among the Δ analyses is generally weaker, except for the lower troposphere of the NH. Intercomparison of the T analyses shows significant systematic differences, though the agreement in their geographical distributions is generally good. This influences the &Ggr; analyses, which show weak agreement particularly in the lower troposphere. The intercomparison of kinematic and thermodynamic fields for the different analyses gives an indication of the degree of agreement among the Q/CP analyses whose calculations involve many highly sensitive quantities. The intercomparison statistics provide us with a quantitative assessment of the systematic differences, spatial coherence, and quantitative agreement between the different analyses for each meteorological field, but they do not provide a measure of the relative accuracy. To provide such a measure, we compare the heating rates and vertical motion with equivalent blackbody emission temperatures observed by TIROS-N in the TR. Through this technique, we infer the relative accuracy of the three FGGE Level IIIb analyses. It is clear that the method of objective analysis for meteorological observations must be refined to improve the quality of the divergent wind analysis and the temperature analysis in the tropics. ¿ American Geophysical Union 1989