A Comparative Study of Observed Northern Hemisphere Circulation Statistics Based on GFDL and NMC Analyses. Part I: The Time-Mean Fields

Abstract

Two sets of monthly mean analyses based essentially on the same observational data, but employing different analysis procedures, are compared. The first set was compiled at the Geophysical Fluid Dynamics Laboratory and consists of horizontal interpolations of monthly averaged circulation statistics accumulated at individual rawinsonde stations. The second set was derived from twice-daily gridded analyses produced by the National Meteorological Center on an operational basis. The data used cover nine winters and nine summers within the 1963–73 period. The spatial domain extends in latitude from 20° to 90°N, and in the vertical from 850 to 100 mb. The circulation statistics examined include 1) hemispheric distributions of 9-year averages as well as month-to-month standard deviations for the horizontal wind components and geopotential heights at 850, 500 and 200 mb, and the temperature at 850 mb; and 2) latitude-height sections for the zonally averaged wind and temperature, the standing eddy variances of zonal and meridional wind components, geopotential height and temperature. and the meridional transports of westerly momentum, geopotential energy and heat by standing waves. Over certain data-sparse regions, the two analyses are further compared with actual values reported in Monthly Climatic Data for the World by the few rawinsonde stations located in those regions. The time-mean fields in the two data sets are found to be generally in excellent merit over the North American and Eurasian continents, where a dense observing network exists. The deviations between the data acts are 1arge over the oceans and northern Africa, where the GFDL analyses give relatively weaker zonal wind speeds in the jet exit regions, stronger ageostrophic motions in the meridional direction, lower temperatures in the subtropical lower troposphere, and higher temperatures above the subtropical tropopause. The maximum local deviations are on the, order of 10–15 m s−1 for zonal wind, 6–m s−1 for meridional wind, 50–70m for geopotential height, and 2–4°C for temperature. These discrepancies are associated with much weaker standing eddy kinetic energy, and much stronger equatorward transports of geopotential energy by the stationary waves in the GFDL analyses. The inter-monthly variability of the monthly mew fields in the GFDL act is generally weaker over the oceans. The spatial correlation coefficients for the monthly mean fields in the two data acts do not exhibit any discernible trends during the 9-year period. This suggests that the procedural changes in the NMC analysis system during this period did not result in serious inhomogeneities in the time series of the NMC fields.