Abstract
Time and zonally averaged grid-point and spectral domain statistics calculated from NMC analyses of wind and temperature which are used as input for the NMC temperate latitudes and tropical operational forecast models have been compared in the overlap zone 2IN to 33N for the periods 00 GMT 1–15 August 1970 and 00 GMT 3–17 January 1971 at 700 mb and 300 mb and have been compared with pertinent general circulation data published by Oort and Rasmusson (1971) for the months of August and January. It was desired to find the best 6° latitude ring within the zone studied in which to merge the two analyses to obtain a horizontally continuous hemispheric data field, with merging to occur in the same latitude ring in the vertical and for both wind and temperature.
Considerable discrepancy between the representation of the same wind and temperature observations by the two analyses is possible at all latitudes, particularly 21N to 30N and in the January period. Due to the discrepancies which were seen, the choice of the best merging-zone in the January period is not as clear-cut as for the August period. Collectively considering the various methods of comparing the two analyses, it appeared that they should be merged from 27N to 33N in both periods.
